WSIS Forum 2026
Rapport généré par l'IA

Empowering Voices: Strengthening CSO and Youth Participation in AI Governance

5 intervenants
Résumé

Résumé

La discussion a porté sur la gouvernance de l'IA, en mettant particulièrement l'accent sur la nécessité de la rendre plus inclusive, démocratique et réceptive à la participation de la société civile et des jeunes . Le modérateur a cadré la session autour d'outils et de mécanismes concrets permettant une implication sociétale plus large dans la gouvernance de l'IA, tandis que les panélistes ont présenté des projets couvrant la documentation communautaire, l'infrastructure d'évaluation et la mise en œuvre réglementaire . Emaan Yasin Malik a indiqué que ses préoccupations ont été suscitées après qu'un hackathon technologique pour les jeunes au Pakistan a attiré des candidatures provenant majoritairement de jeunes hommes musulmans, avec une faible participation féminine et aucune représentation des communautés hindoue, chrétienne ou sikhe. Des questions se sont également posées quant à savoir si la société civile avait été véritablement consultée lors de l'élaboration de la politique nationale pakistanaise en matière d'IA . En réponse, Malik a décrit un projet pilote mené avec la communauté chrétienne visant à documenter les préjudices facilités par l'IA, tels que les deepfakes entraînant des accusations de blasphème et des violences. Parmi les objectifs supplémentaires figurent la création de voies d'orientation vers le système de plaintes de la Commission nationale des droits de l'homme, ainsi que l'élaboration de ressources dans les langues locales, notamment l'ourdou et le pendjabi . Addie Achan a ajouté que le projet Kaleidoscope aide les communautés locales à évaluer et à gouverner l'IA. Elle a exprimé sa crainte que les projets « IA pour le bien commun » n'exportent des biais californiens ou américains dans d'autres contextes sans garde-fous appropriés . Vasileios Rovilos a expliqué que Credo AI se concentre sur la couche opérationnelle de la réglementation en constituant des bibliothèques de risques, des bibliothèques de contrôles et des techniques d'atténuation afin de créer des structures de preuve et de responsabilité pour la gouvernance de l'IA . Un thème majeur a été la nécessité pour la société civile de disposer d'un meilleur accès aux données probantes, aux financements et à la collaboration. Malik a soutenu qu'en Asie du Sud, les OSC peinent à construire des infrastructures de preuve indépendantes, car leur travail est financé par des donateurs, à court terme, répétitif et cloisonné, tandis que l'accès aux ensembles de données détenus par les gouvernements est également difficile . Achan a abondé dans ce sens en soulignant que beaucoup de travaux se déroulent en vase clos, et a suggéré que les OSC peuvent collecter des données probantes qui sont ensuite codifiées via des infrastructures existantes, citant en exemple une fiducie de données en langue māorie en Nouvelle-Zélande comme modèle de gestion communautaire . Rovilos a plaidé pour la collaboration, l'intégrité des données probantes et des taxonomies communes afin d'éviter la fragmentation dans la gouvernance mondiale de l'IA . Sur la question de la participation des jeunes, les panélistes ont convenu que les jeunes ont besoin d'un pouvoir réel plutôt que d'une inclusion symbolique. Malik a souligné les inégalités d'accès parmi les étudiants, les barrières linguistiques dans l'explication de l'IA, la croyance erronée selon laquelle une expertise en programmation est nécessaire, ainsi que la valeur des universités, des forums de jeunesse et des institutions nationales des droits de l'homme dans la construction d'un engagement en matière d'IA fondé sur les droits . Achan a distingué le tokenisme d'une véritable prise de décision partagée, en soutenant que les jeunes devraient avoir autorité sur les décisions qui les concernent . Rovilos a ajouté que les jeunes figurent parmi les personnes les plus exposées aux systèmes d'IA, des outils de surveillance des examens aux outils de recrutement, et devraient donc être impliqués dès le début, se voir accorder une influence et être soutenus par des actions concrètes et des efforts en matière de culture numérique de l'IA . Dans une perspective à cinq ans, les intervenants ont appelé à des rôles plus formalisés et mieux dotés en ressources pour les jeunes dans la gouvernance de l'IA. Malik a proposé une participation obligatoire des jeunes aux processus internationaux de reporting numérique et la création d'une commission numérique indépendante dirigée par des jeunes au Pakistan . Achan a plaidé pour un vote participatif transparent sur les priorités en matière d'IA, associé à des évaluations indépendantes permettant de vérifier si les systèmes répondent aux besoins du public . Dans l'échange final, le panel a également noté que les parents et les enseignants ne sont pas encore pleinement équipés pour enseigner l'IA, mais pourraient faire davantage avec des ressources accessibles, tandis que la gouvernance des systèmes d'IA plus avancés devrait progresser à travers des attentes communes, des processus multipartites et un renforcement de la culture numérique de l'IA . Dans l'ensemble, la discussion a convergé vers la collaboration, les données probantes et la participation des jeunes comme conditions essentielles d'une gouvernance de l'IA plus équitable .

Points clés

- Le thème central du panel était la construction d'une gouvernance de l'IA plus inclusive, plurielle et démocratique, en impliquant la société civile, les communautés locales et les jeunes, le modérateur ayant cadré la session autour d'outils et de mécanismes concrets pour une participation plus large à la gouvernance de l'IA. - Emaan Yasin Malik a mis en lumière l'exclusion et la sous-représentation dans le paysage technologique et politique de l'IA au Pakistan, notant une participation déséquilibrée lors d'un hackathon pour les jeunes et des préoccupations selon lesquelles la politique nationale pakistanaise en matière d'IA affirmait avoir consulté des groupes de défense des droits, ce que beaucoup ont démenti. En réponse, elle a décrit un projet pilote axé sur la communauté chrétienne visant à documenter les préjudices facilités par l'IA, à créer des voies d'orientation vers la Commission nationale des droits de l'homme et à élaborer des ressources communautaires dans les langues locales. - Addie Achan et Vasileios Rovilos ont tous deux souligné que l'efficacité de la gouvernance de l'IA repose sur des données probantes et une infrastructure opérationnelle. Addie a décrit l'approche de Kaleidoscope comme « données probantes plus infrastructure égale confiance », en utilisant les contributions des communautés pour élaborer des référentiels et des évaluations de l'IA pour les déploiements locaux. Vasileios a de même plaidé pour des bibliothèques de risques, des bibliothèques de contrôles et des techniques d'atténuation qui traduisent la réglementation en pratique opérationnelle et créent des structures de responsabilité. - Un point de discussion majeur a été la difficulté que rencontrent les organisations de la société civile à participer de manière significative à la gouvernance de l'IA, en particulier dans les contextes à faibles ressources. Emaan a évoqué les cycles de financement courts imposés par les donateurs, la duplication des projets, le travail cloisonné et les difficultés d'accès aux ensembles de données gouvernementales fragmentées en Asie du Sud. Addie a plaidé pour la collaboration et pour que les OSC collectent des données probantes pouvant être codifiées via des structures de gouvernance existantes, tandis que Vasileios a insisté sur la combinaison de données probantes indépendantes, de taxonomies communes et de collaboration pour éviter la fragmentation et accroître l'influence. - La participation des jeunes est apparue comme une priorité essentielle, les intervenants soutenant que les jeunes ont besoin de plus qu'une représentation symbolique. Emaan a souligné les inégalités d'accès, les barrières linguistiques et les idées reçues selon lesquelles seuls les programmeurs peuvent façonner la gouvernance de l'IA, et a proposé de travailler à travers les écoles, les universités et les institutions nationales des droits de l'homme. Addie a utilisé un cadre opposant le tokenisme à une véritable prise de décision partagée, en soutenant que les jeunes devraient disposer d'une autorité réelle. Vasileios a renforcé la nécessité d'une implication proactive, d'un suivi concret et d'une culture numérique de l'IA, notamment parce que les jeunes figurent parmi les personnes les plus touchées par les systèmes d'IA. Dans une perspective à cinq ans, le panel a appelé à une inclusion obligatoire des jeunes dans les rapports internationaux, à la création d'organes numériques indépendants dirigés par des jeunes, à des systèmes participatifs transparents pour l'évaluation des outils d'IA, à davantage de financements et à des engagements institutionnels plus solides en faveur de la culture numérique de l'IA.

L'objectif général de la discussion était d'explorer comment la gouvernance de l'IA peut devenir plus inclusive, fondée sur des données probantes et démocratique, notamment en renforçant le rôle de la société civile et en veillant à ce que les jeunes et les communautés marginalisées ne soient pas simplement consultés de manière symbolique, mais soient en mesure de façonner les décisions dans la pratique.

Le ton général de la discussion a été constructif, réflexif et collaboratif tout au long. Il a débuté de manière pratique et conversationnelle, le modérateur invitant explicitement à la discussion plutôt qu'à des présentations formelles. Il est ensuite devenu plus sérieux et urgent lorsque les intervenants ont décrit l'exclusion, la faiblesse des consultations, les préjudices en ligne et les obstacles structurels à la participation. Vers la fin, le ton est devenu plus prospectif et orienté vers les solutions, notamment lors de la discussion sur la participation des jeunes, les futures réformes institutionnelles et les questions du public sur l'AGI, l'éducation et le dialogue intergénérationnel.

Intervenants

- Dr. Pedro dos Santos Maia - Modérateur de la session sur la gouvernance de l'IA et la gouvernance inclusive et diversifiée de l'IA ; a présidé la discussion et la séance de questions-réponses. - Mme Emaan Yasin Malik - Praticienne des droits de l'homme spécialisée dans les droits numériques, la protection des femmes et des enfants, et les droits des communautés minoritaires ; coordinatrice de programme à la Commission/Comité national des droits de l'homme du Pakistan. - Mme Addie Achan - Fondatrice et directrice exécutive du projet Kaleidoscope, une organisation à but non lucratif axée sur la réduction des risques liés à l'IA, la collecte de données probantes et l'implication des communautés locales dans le mouvement IA pour le bien commun ; ancienne ingénieure en IA et ancienne responsable IA au sein d'une fondation philanthropique finançant des déploiements d'IA pour le bien commun. - M. Vasileios Rovilos - Directeur de la politique européenne chez Credo AI ; dirige la pratique de politique européenne de l'entreprise et supervise la mise en œuvre de l'AI Act de l'UE et des réglementations numériques connexes, notamment le DSA, le DMA, le RGPD, NIS2 et le CRA. - Public - Participants du public posant des questions lors de la séance de questions-réponses. Intervenants supplémentaires : - Aucun clairement identifié par son nom au-delà des intervenants listés.

Intervenants
ME
Ms. Emaan Yasin Malik
163 wpm · 11 min
MA
Ms. Addie Achan
178 wpm · 7 min
MV
Mr. Vasileios Rovilos
167 wpm · 16 min
DP
Dr. Pedro dos Santos Maia
155 wpm · 6 min
A
Audience
156 wpm · 41 s

Le modérateur, Dr Pedro dos Santos Maia, a présenté le panel comme une conversation sur la manière de construire une gouvernance de l'IA plus inclusive et diversifiée, en l'articulant explicitement autour de mécanismes pratiques plutôt que de discours formels . Il a invité les trois panélistes - Emaan Yasin Malik, Addie Achan et Vasileios Rovilos - à expliquer quels outils ou processus concrets ils développaient, quelle lacune en matière de gouvernance ces outils comblaient, et pourquoi ils avaient choisi ce point de départ . Cela a donné à la discussion une orientation pratique axée sur des mécanismes concrets d'inclusion dans la gouvernance de l'IA .

Malik a ancré ses remarques liminaires dans le contexte institutionnel et social du Pakistan . Elle a expliqué que la Commission nationale des droits de l'homme se situe entre la société civile et le gouvernement, et cherche à faire le lien entre les deux . Elle a ensuite décrit un hackathon technologique pour les jeunes organisé avec la Cour suprême, dans lequel la majorité des candidats étaient de jeunes hommes musulmans, environ 20 pour cent seulement étaient des femmes, et aucun candidat ne provenait des communautés hindoue, chrétienne ou sikh . Elle a présenté cela comme le signe d'une exclusion plus large, façonnée par des hiérarchies sociales et communautaires . Elle a établi un lien avec les préoccupations concernant la politique nationale d'IA du Pakistan, indiquant que bien que cette politique affirme que la société civile et les groupes de défense des droits avaient été consultés, plusieurs organisations de premier plan ont déclaré ne pas avoir été réellement incluses .

Malik a ensuite présenté un projet pilote centré sur la communauté chrétienne du Pakistan . La première étape consistait à documenter les préjudices liés à l'IA avec les églises, les écoles chrétiennes, les OSC et les membres de la communauté . Elle a donné l'exemple des deepfakes susceptibles de déclencher des accusations de blasphème et, par conséquent, d'entraîner des incitations à la violence, des violences de foule ou des arrestations . La deuxième étape visait à créer une voie d'orientation vers le système de plaintes de la Commission, afin que les agents chargés des plaintes puissent traiter ces cas plus efficacement . La troisième étape, qu'elle a décrite en termes plus prudents, consistait à développer des ressources accessibles en ourdou et en pendjabi, et à veiller à ce que les contributions de la communauté parviennent également aux processus d'élaboration des politiques .

Achan a présenté l'approche de Kaleidoscope centrée sur la communauté en matière de gouvernance de l'IA . S'appuyant sur son expérience d'ancienne ingénieure en IA, puis de responsable IA au sein d'une fondation philanthropique soutenant des travaux d'IA à vocation sociale, elle a dit croire toujours au potentiel bénéfique de l'IA, mais a soutenu que de nombreux déploiements manquaient de garde-fous adéquats . Sa préoccupation était que des outils largement développés en Californie ou aux États-Unis pouvaient être exportés dans d'autres contextes en emportant avec eux des présupposés et des biais centrés sur les États-Unis, même lorsqu'ils étaient présentés comme socialement bénéfiques . Elle a résumé son approche par la formule : « preuve plus infrastructure égale confiance » . Par preuve, elle entendait la collecte des contributions des communautés ; par infrastructure, elle entendait la transformation de ces contributions en outils de gouvernance utilisables, tels que des référentiels et des évaluations de l'IA, qu'elle a décrits comme des tests permettant de comprendre les performances des systèmes . Son argument était que la confiance est plus susceptible d'émerger lorsque les communautés contribuent à façonner les normes selon lesquelles les systèmes d'IA sont évalués .

Rovilos s'est concentré sur le volet opérationnel de la gouvernance . S'exprimant au nom de Credo AI, il a indiqué que l'organisation travaille sur la couche qui traduit les exigences réglementaires en pratiques quotidiennes . Il a vivement approuvé la position d'Achan selon laquelle la preuve deviendrait centrale dans la gouvernance de l'IA, et a dit que cela serait également important pour permettre la participation de la société civile et des jeunes . En même temps, il a noté que les OSC sont souvent invitées à regarder « sous le capot » des systèmes d'IA sans disposer de preuves suffisantes pour le faire . Il a indiqué que Credo AI répond à cela en constituant des bibliothèques de risques, des bibliothèques de contrôles et des techniques d'atténuation, et en les élaborant avec ses clients afin que les organisations disposent d'une base de preuves utilisable et d'une piste d'audit pour la responsabilisation .

Le modérateur a ensuite dégagé un fil conducteur commun aux trois interventions : la relation entre la preuve et l'infrastructure de gouvernance . Il a demandé si les OSC avaient besoin d'une infrastructure de preuve indépendante qui leur soit propre pour participer de manière significative à des systèmes souvent dominés par des experts et des ingénieurs, ou si les connaissances de la société civile et l'expérience vécue pouvaient plutôt être intégrées dans les cadres de gouvernance existants .

Malik a répondu que la réponse dépendait fortement du contexte national, en particulier en Asie du Sud . Elle a dit que les OSC peinent souvent à construire leurs propres systèmes de preuve, car ce travail est financé par des donateurs, les cycles de financement sont courts, et la concurrence pour les subventions favorise la duplication et les projets cloisonnés . Elle a ajouté que les organisations finissent fréquemment par répéter des formes similaires de collecte de données sans créer d'infrastructure partagée durable . Elle a soutenu qu'une plus grande collaboration entre les OSC était nécessaire, éventuellement avec le soutien d'un organisme distinct . Pour illustrer le problème des ressources, elle a décrit des bénévoles travaillant sur le Forum de gouvernance de l'Internet Asie-Pacifique au Pakistan après leurs heures universitaires, tout en devant encore chercher des donateurs pour organiser le forum annuel . Elle a également souligné les difficultés à obtenir des données, notant que dans de nombreux pays d'Asie du Sud, les ensembles de données clés sont fragmentés entre les institutions . Son exemple était la tentative de l'UNICEF au Pakistan de développer un outil de vérification de l'âge basé sur l'IA pour les enfants avant leur transfert vers des quartiers de prison pour adultes, entravée par des problèmes d'obtention de données policières auprès du Bureau national de la police .

Achan a globalement convenu que trop de travaux se déroulent en vase clos . Elle a soutenu que les OSC peuvent rassembler des preuves, mais que ces preuves doivent ensuite se connecter à des structures institutionnelles pour devenir exploitables . Elle a illustré cela par un exemple de Nouvelle-Zélande, où des communautés māories développaient la reconnaissance vocale pour une langue sous-dotée en ressources . Lorsqu'une société cinématographique américaine a tenté d'acheter les données, une OSC indépendante a aidé à les organiser, et le gouvernement les a finalement placées dans un fonds fiduciaire afin qu'elles puissent bénéficier à la communauté et être protégées contre la surveillance ou une utilisation commerciale extractive . Elle a dit que cela montrait que les OSC peuvent collecter et organiser des preuves communautaires, tandis que les gouvernements peuvent fournir des garanties juridiques ou institutionnelles autour de celles-ci . Elle a également élargi la discussion, passant de l'accès aux preuves à la question de ce qui compte comme preuve dans la gouvernance de l'IA et de la manière dont les OSC peuvent y contribuer .

Rovilos a dit qu'il ne fallait pas traiter cela comme un choix binaire entre preuve et infrastructure . Il a plutôt plaidé pour les deux : l'indépendance des OSC peut contribuer à préserver l'intégrité des preuves, tandis que des taxonomies partagées et un langage commun peuvent permettre à ces preuves de circuler entre les institutions et les juridictions . Sans de tels cadres communs, a-t-il averti, la gouvernance risque de devenir une « Tour de Babel » dans laquelle des acteurs fragmentés ne peuvent pas communiquer de manière significative . Il a également cité des chiffres suggérant que si une grande majorité d'entreprises utilisent désormais l'IA, une très faible proportion sait comment mettre en œuvre la gouvernance de l'IA en pratique . Il a utilisé cela pour souligner à la fois le fossé de mise en œuvre et la nécessité d'une collaboration entre les acteurs qui cherchent à construire des systèmes fonctionnels .

La conversation s'est ensuite tournée vers la participation des jeunes . Le modérateur n'a pas simplement demandé comment inclure les jeunes ; il a demandé s'ils devaient être impliqués de la même manière que les experts actuellement, ou si quelque chose de plus spécifique était nécessaire . Cette question a orienté le tour de parole suivant.

Malik a mis l'accent sur les inégalités d'accès, de langue et de confiance en soi . Elle a mis en contraste les élèves des écoles d'élite, qui peuvent avoir un meilleur accès aux outils, aux modèles de langage et aux cours en ligne, avec les élèves des écoles publiques sous-dotées en ressources, où même expliquer l'IA en ourdou peut s'avérer difficile . Elle a raconté avoir essayé de traduire l'IA par « nakli akal », que les élèves ont compris de manière négative et ont trouvé alarmant . Elle a également dit que de nombreux jeunes associent principalement l'IA aux outils d'image des réseaux sociaux et supposent qu'ils doivent savoir coder avant de pouvoir parler de gouvernance de l'IA . Pour elle, cela signifiait que la participation significative dépend autant de la littératie et de l'accessibilité linguistique que de l'inclusion formelle . Elle a mentionné des espaces régionaux anglophones tels que Netmission Asia et le Forum de gouvernance de l'Internet pour les jeunes de l'Asie-Pacifique . Au Pakistan, cependant, elle a dit que les universités constituaient l'un des points d'entrée les plus utiles, et elle a évoqué des mémorandums d'accord avec 20 universités et des plans pour travailler à travers des cellules de droits de l'homme afin de développer des modules sur l'IA et les droits . Elle a également soutenu que les institutions nationales des droits de l'homme devraient disposer d'un programme ou d'un point focal sur l'IA et les droits .

Achan a abordé la participation des jeunes en remettant en question les présupposés sur l'expertise . Elle a soutenu que, l'IA étant encore relativement nouvelle, la différence d'utilisation de l'IA entre une personne de 25 ans et une personne de 50 ans n'est peut-être pas si marquée . Elle a également noté que les conséquences à long terme de l'IA, y compris les effets environnementaux, affecteront particulièrement les jeunes générations . Pour cadrer son propos, elle s'est appuyée sur l'échelle de participation de Roger Hart, qui va du tokenisme au bas de l'échelle à la prise de décision partagée initiée par les jeunes au sommet . Son argument central était que la gouvernance ne devrait pas simplement « donner la parole aux enfants » sans leur offrir des choix ou une influence significatifs . Elle a suggéré d'évaluer la participation en se demandant si la présence des jeunes modifie réellement le résultat . Cela a renforcé la distinction entre l'inclusion symbolique et le véritable pouvoir de décision pour les jeunes .

Rovilos a repris l'argument anti-tokenisme mais y a ajouté une critique institutionnelle plus forte . Il a dit que la tokenisation est un moyen peu coûteux de paraître inclusif tout en refusant le vrai pouvoir . Selon lui, la participation des jeunes doit être liée à des processus qui comptent, et la distance structurelle entre les jeunes et les décideurs doit être abordée directement . Il a soutenu la position de Malik sur les ressources et le suivi, en faisant valoir que les jeunes ont besoin de soutien, d'un engagement soutenu et d'une influence réelle plutôt que d'invitations ponctuelles . Il a également dit que la participation des jeunes devrait être liée aux systèmes d'IA déployés localement, car la proximité avec les préjudices vécus est importante . Le calendrier était également important pour lui : les jeunes devraient être impliqués tôt, avant que les décisions ne soient effectivement arrêtées . Il a ajouté que ce n'est pas seulement un point de procédure, car les jeunes sont déjà fortement affectés par les systèmes d'IA à travers des outils tels que les logiciels de surveillance d'examens, les algorithmes des téléphones et, plus tard, les systèmes de recrutement . D'un point de vue européen, il a mentionné le Pacte sur l'IA comme un possible véhicule multipartite, mais a soutenu que les efforts actuels en matière de littératie de l'IA sont principalement conçus pour les déployeurs et les acteurs corporatifs, laissant les jeunes largement en dehors du processus .

Dans la section prospective, le modérateur a demandé quelles institutions, quels cadres ou quels référentiels pour la participation des jeunes les panélistes aimeraient voir dans cinq ans .

Malik a proposé deux changements principaux . Premièrement, elle a dit que la participation des jeunes devrait être obligatoire dans les rapports internationaux sur les questions numériques, en particulier dans les processus liés à l'ONU, afin que les États soient tenus de produire des preuves de ce que les jeunes disent réellement . Elle a noté que les enfants sont déjà consultés dans le cadre de la Convention relative aux droits de l'enfant et a suggéré que les rapports liés au numérique dans le cadre de mécanismes tels que le CAT et le CERD devraient également inclure les contributions des jeunes . Deuxièmement, elle a dit que dans cinq ans, elle aimerait voir une commission numérique indépendante pour les jeunes au Pakistan, gérée par les jeunes eux-mêmes mais financée par l'État . Elle a présenté cela comme une aspiration liée aux préoccupations concernant la liberté d'expression et les violences généralisées en ligne et facilitées par la technologie affectant les enfants, les communautés minoritaires et les femmes .

Achan a dit qu'elle aimerait voir une forme de processus public qui interroge continuellement les communautés sur ce qu'elles attendent de l'IA, potentiellement à travers des mécanismes spécifiques à des enjeux particuliers plutôt qu'une structure unique à tout faire . Elle a ensuite proposé un système de vote participatif transparent permettant aux personnes d'exprimer ce qu'elles attendaient des systèmes d'IA entrant dans leur vie . Ces préférences seraient mises en correspondance avec des scores transparents montrant dans quelle mesure ces systèmes répondaient réellement aux besoins des communautés . Elle a souligné que cette notation devrait être réalisée par un évaluateur indépendant plutôt que par les entreprises d'IA elles-mêmes .

Rovilos a globalement soutenu les propositions précédentes et y a ajouté trois priorités . Premièrement, il a dit que la gouvernance de l'IA est intrinsèquement intergénérationnelle, de sorte qu'exclure les jeunes aurait peu de sens étant donné qu'ils vivront le plus longtemps avec ses conséquences . Deuxièmement, il est revenu sur la question du financement et a soutenu que davantage de ressources doivent être investies dans la participation des jeunes au cours des cinq prochaines années . Troisièmement, il a critiqué la pratique européenne qui consiste souvent à inviter les jeunes à soulever des préoccupations sans véritable suivi . Ce qu'il souhaitait à la place, c'était une inclusion davantage axée sur l'impact, avec des droits de participation plus clairement accentués, voire renforcés, et des institutions tenues d'agir sur ce que disent les jeunes . Il a conclu en revenant sur la littératie de l'IA, en soutenant que les institutions devraient traiter le renforcement des compétences des jeunes comme faisant partie de leur pratique quotidienne, car les jeunes générations devront faire face aux conséquences à long terme de l'utilisation généralisée de l'IA .

Les questions du public ont introduit deux dernières questions : l'IA avancée et l'éducation intergénérationnelle . Un membre du public a fait référence à une déclaration attribuée à un dirigeant de Google selon laquelle l'AGI pourrait arriver avant 2030, et a demandé ce que cela signifierait pour la gouvernance de l'IA et si les jeunes seraient impliqués . Un autre a demandé comment le dialogue intergénérationnel pourrait être construit en pratique et si les parents et les enseignants étaient équipés pour enseigner l'IA à la prochaine génération .

Achan a répondu que les parents et les enseignants ne sont pas encore équipés pour bien enseigner l'IA, car c'est un domaine nouveau et l'expertise en la matière n'est pas étroitement liée à l'âge . Elle a dit que cela renforce en fait l'argument en faveur d'une voix égale pour les jeunes, plutôt que de supposer que les générations plus âgées devraient décider à leur place . Sur l'AGI, elle a adopté une vision relativement continue, soutenant que même si les systèmes deviennent beaucoup plus capables, la gouvernance devra toujours se concentrer sur la manière dont ils affectent l'expérience humaine . Pour cette raison, les jeunes resteraient des participants essentiels, bien que non exclusifs, au sein d'une gouvernance intergénérationnelle .

Malik a répondu à la question sur l'éducation avec des exemples du Pakistan . Elle a décrit les enseignants comme surchargés de travail mais généralement ouverts aux outils susceptibles de les aider . Elle a également observé que les parents sont souvent anxieux mais très motivés, donnant l'exemple de parents à Lahore demandant comment « faire entrer leur enfant dans l'IA » . Malgré la confusion et les récits médiatiques contradictoires, elle a dit que cela crée un certain degré de volonté et d'enthousiasme chez les parents, les enseignants et les élèves . Sa recommandation pratique était de fournir des programmes et des ressources d'ateliers gratuits, accessibles et adaptés aux langues locales, car les enseignants sont susceptibles de les utiliser s'ils sont disponibles .

Rovilos a répondu d'un point de vue réglementaire et institutionnel . Sur l'AGI, il a reconnu que des avancées majeures pourraient survenir plus rapidement que prévu, mais a soutenu que la gouvernance devrait néanmoins progresser étape par étape . Il a cité la loi européenne sur l'IA comme exemple d'un cadre qui, au moins, classe les risques et y attache des obligations, y compris des pratiques interdites . Sa préoccupation plus profonde était que des systèmes plus avancés pourraient remettre en question des présupposés tels que la supervision humaine significative . Il a également mentionné des dynamiques géopolitiques plus larges, notamment les contrôles à l'exportation et l'accès aux technologies, comme preuve que les États façonnent déjà les marchés de l'IA de manière importante . En même temps, il a soutenu que la priorité immédiate est d'établir des attentes communes à travers un large engagement multipartite plutôt que d'essayer de résoudre tous les scénarios futurs à la fois . Sur l'éducation, il a donné l'exemple de sa nièce utilisant Claude sur une tablette pour ses devoirs à la maison, à l'insu de ses enseignants . Il a directement lié cela à la nécessité d'obligations en matière de littératie de l'IA non seulement pour les entreprises, mais aussi pour les écoles et les institutions publiques .

Dans l'ensemble, le panel a été marqué davantage par des complémentarités que par des désaccords . Malik s'est concentrée sur l'exclusion, les préjudices subis par les minorités, les barrières linguistiques, les voies de recours et la nécessité de mandats institutionnels . Achan a mis l'accent sur les preuves communautaires, la confiance, l'évaluation indépendante et la différence entre la participation symbolique et l'influence réelle . Rovilos s'est concentré sur l'opérationnalisation de la gouvernance, la construction de cadres communs, l'évitement du tokenisme et le fait que les efforts actuels en matière de littératie sont trop axés sur les déployeurs et les acteurs corporatifs . À travers ces contributions, plusieurs thèmes communs ont émergé : l'inclusion nécessite des structures pratiques ; la preuve et l'infrastructure doivent fonctionner ensemble ; et la participation est affaiblie lorsque le financement est instable, que les données sont inaccessibles ou que les institutions ne donnent pas suite .

En conclusion, la discussion a suggéré un modèle stratifié de gouvernance inclusive de l'IA ancré dans les préjudices vécus, les contributions des communautés et des mécanismes institutionnels utilisables . Elle a également établi un lien entre la pertinence locale et les exemples pakistanais de Malik d'une part, et l'avertissement d'Achan concernant les biais américains importés d'autre part . Le panel s'est terminé sans résolutions formelles, mais avec une orientation claire : une gouvernance de l'IA plus équitable nécessitera des preuves ascendantes, une collaboration renforcée, une participation significative des jeunes, et des institutions capables de traduire la participation en responsabilisation et en pratique . Le modérateur a conclu en remerciant Globethics d'avoir organisé la session et en invitant les participants à poursuivre la conversation de manière informelle par la suite .

Dr. Pedro dos Santos Maia
Can you hear me? No. Hello. I think it's kind of like the volume is low or something. Oh. Oh, okay. Hello. Hello. Hello. Hello. Hello. Hello. Hello. Hello. Bonjour. Ah, okay. Can you hear me now with the headphones? I think so, right? Awesome. Thank you so much for the feedback. So, hi, everybody. Nice to see you all here for this session. I didn't want to use the gavel. I thought it was going to be a little bit hardcore, so thank you. So today we have Addie, we have Emaan and Vasileios. We're going to talk about AI governance and how to be able to build a more inclusive and diverse governance of AI. So I'm going to start by introducing our fellow members of the table. We have Addie Chan. She's the founder and executive director of the Kaleidoscope Project, a nonprofit dedicated to the risking, curating evidence for and involving local communities in the AI for Good movement. We also are joined by Emaan. She's a human rights practitioner focusing on digital rights, the protection of women and children and the rights of minority communities. She's now been working as a program coordinator at Pakistan's National Committee for Human Rights. And last but not least. OK. Thank you. And if we get closer to the microphone, last but not least, we have Vasileios. He's the EU Policy Director at Credo AI, where he has the EU policy practice and directs the company's implementation of the EU AI Act and adjacent digital regulations such as DSA, DMA, GDPR, NIS2, and CRA. So thank you all for joining us. And the idea of today's session is to be more of a conversation and a discussion. And we separated a few questions to kind of get the ball rolling. And I'd like to start with a more general question. And the idea is to kind of ask you here at the table, we know you've been working with AI, you've been worried about AI governance for the last year. for the last years, and we wanted to hear more from you and understand what are you building in terms of concrete tools and mechanisms to ensure a more plural and democratic participation of civil society and also other parts of society in AI governance. So please tell us what are you developing, what gap it fills in terms of AI governance, and why you chose to start there. We can go from left to right. And please feel free to start the floor. Thank you.
Ms. Emaan Yasin Malik
Perfect. Hi. I'll speak a bit louder because I'm not sure if you guys can hear in the headphones. So I work at Pakistan's National Commission for Human Rights, which is an independent body, which means that we're not exactly a CSO or government. We sit somewhere in the middle and we try to bridge these two bodies. So the problem that I am tackling is. I'm not sure if you guys can hear me. started actually when our organization hosted a tech hackathon for youth in Pakistan. So we were working in collaboration with the Supreme Court. And what we realized was that all of our applications were from young Muslim men. And we had a few women, 20 % women applications, and no applications from the Hindu community, Christian community, Sikh community. So within Pakistan, these are kind of the class structures that exist. And this is something that the commission was very concerned with. The second concerning factor was the introduction of Pakistan's national AI policy, which came out last year. And although the policy claimed that they had consulted CSOs, rights organizations, human rights bodies, a lot of the leading organizations said that they were not consulted in this process. So we were very concerned. And with that said, the presentation is now over. The project that I'm kind of working on is that we are using as a pilot. We are working with the Christian community, starting with a data collection and documentation process where we are hoping to work with the church, Christian high schools, middle schools, community members and CSOs to kind of document what harms they are facing in terms of AI facilitated violence, which includes deepfakes leading to blasphemy allegations, which can lead to incitement, mob violence, arrests and so we are really starting at the ground level with children. After that, we are looking to form a referral pathway leading to the commission's complaint system, hoping that our complaints officers are equipped and ready to deal with these complaints. And then after that, we are also hoping to create different resources for these communities in their own languages, Urdu, Punjabi, and kind of creating a holistic system where their input is available in the policies that are being shaped, where they have a referral pathway, a complaints mechanism, et cetera. So I'll move on the conversation to my colleague.
Dr. Pedro dos Santos Maia
Thank you so much, Emaan. I'm also going to follow your cue and also going to speak louder. Now I'll pass the word to Addie. Thank you so much.
Ms. Addie Achan
Hey, everyone. So nice to meet you. I'm Addie Chan. I'm from the United States. And as mentioned in my intro, I founded an organization called Kaleidoscope. Essentially, we're focused on helping local communities both evaluate and govern AI as they enter into them. A little bit about my background is that I used to be an AI engineer, and then later I was the AI lead at a philanthropy that allocated a lot of AI for good deployments. And while I think AI for good is like obviously a feasible thing, I also think the movement as a whole may be approaching it without the necessary guardrails, especially as they think about importing what's a technology that, as I live in California, is largely developed in California or U .S. biases and then deployed by people in the U .S. and these other spaces in the name for good. So essentially, I try to like operate under this equation, which is basically evidence plus infrastructure equals trust. And evidence means collecting community inputs. Infrastructure is converting that into necessary tools like benchmarks or AI evaluations, which, in case someone's not familiar, is just a test to understand the performance of an AI. And the hope is that. when you center those in deployment, then you can gain back community's trust rather than currently the feeling of it being forced on them. Yeah, thank you.
Dr. Pedro dos Santos Maia
Awesome. Addie, thank you so much. Now I'm going to pass the word to Vasileios, please, before it's yours.
Mr. Vasileios Rovilos
Yeah, thank you so much. And I'm going to make sure to also project. I think that was, like, fantastic to actually hear, you know, like, more, like, insightful, like, the elements that you actually do on a day -to -day. I think that's truly fantastic. Like, what Credo AI does in that direction is that basically we try to offer the operational layer of what actually happens in the regulatory space and sort of, like, translate that into, you know, the elements that one way or another make sense to do operationally, things that you need to live up to, and one way or another make sure to actually, you know, have a more robust and more forward -leaning and more sustainable and forward -leaning as well, you know, approach in AI governance. More holistic, in essence. So I think that's true. truly fantastic to create a panoply of resources. Very much to your point, Addie, evidence is basically what is going to be differentiating AI governance down the line and a key enabler for civil society organizations, for youth participation. And it actually feels that CSOs, more broadly, are having the ex officio task to actually look under the hood. And yet again, they don't have too much evidence to work with to begin with. And one way or another, that's a major problem to try to elucidate, one way or another. And it's not enough talk about it. And that's quite sad. So one way or another, what Create .ai does is create a risk library, controls library, risk mitigation techniques, co -shape those with our customers to make sure that one way or another, there's this very concrete, substantive body, inventory, if you will, of evidence that one way or another actually puts the trail down. And that's a very important thing to do. And it's not enough to just say, hey, we're going to do this. We're going to do this. We're going to do this. We're going to do this. We're going to do this. We're going to do this. We're going to do this. We're going to do this. We're going to do this. We're going to do this. We're going to do this. and from that point on, accountability structures, the next frontier when it comes to AI governance.
Dr. Pedro dos Santos Maia
Vasileios, thank you so much. So there's a common thread, I think, connecting all of your projects, and I would say they discuss how evidence is produced, how evidence is part of AI governance infrastructures, but there's also in our current stage of AI governance where we have expert groups, high -level groups of not only engineers but also scientists. I was wondering how could CSOs participate in this more actively? Do you think there is a need for some kind of independent evidence infrastructure to shape the decision -making, or there's another alternative for CSOs and their forms of knowledge and their experiences to be translated into current AI frameworks? And I will start with Emaan, then I'm going to move the same way. Thank you, Emaan.
Ms. Emaan Yasin Malik
Okay, great. So first, I guess this question would differ based on where we are, like what the country systems are. Okay. But maybe speaking from like a South Asian perspective or maybe also Asia Pacific in that regard, is that if a CSO, it's very difficult for CSOs firstly to build their own evidence infrastructures. In a country like Pakistan, this work is almost entirely donor funded and they have really short cycles of funding. It's really competitive and this also leads to repeated projects and common, like overarching kind of work. Not overarching, but I would say that a lot of their projects are similar. So they're repeating data collection activities and they're working in silos. So I think the biggest problem is that there needs to be more collaboration between the CSOs and perhaps there can be a body that assists the CSOs in doing this. Of course, I think the donor challenge that you mentioned about finance is a big problem, right? So I work a lot with you. CSOs and youth bodies. I do work with Asia -Pakistan, Asia -Pacific Internet Governance Forum in Pakistan. So we do have a group of 100 volunteers who are working completely free after university trying to get these questions of like AI governance, digital rights policies out. And right now, for example, they're organizing their annual forum, but they're just running around looking for donors. So I think that is a big problem. Another problem is, of course, data access. In a lot of South Asian countries, there are different government bodies that look after different data sets. So, for example, birth registries will look at birth data, police centers will look at police data, and there's a lot of, it's difficult to get access to these systems. So, for example, if UNICEF is looking to develop in Pakistan an AI -powered system to verify the age of children before they are committed to a senior ward in prison, it's very difficult for them to even get that data from the National Police Bureau. So that's a gap that I kind of had in my video.
Dr. Pedro dos Santos Maia
Thank you so much, Emaan. That's also a lot of food for thought. Now I'll pass the word to Addie.
Ms. Addie Achan
Sure. Yeah, I mean, I think Emaan made a great point around collaboration. I think right now there is so much happening in isolation. And a big thing that I think is that I do feel like CSOs have the opportunity to collect evidence and then you can plug into existing infrastructure as like a vehicle to make that convert into action. An example of like a story I've heard is that in New Zealand there's a group of Maori people who were essentially building their own speech recognition for their local language, which is an under -resourced language. And a U .S. company, a film company of everything, tried to buy that data off of them. But an independent CSO was the one that really curated the actual data. But then the government in New Zealand put it in a trust in which that data only benefits the people and it's meant. And it, like, bars it from surveillance or other types of, like, capitalistic gains from it. So that's where I think it's a good demonstration of, like, you can have a CSO collect evidence, and then you can have existing infrastructure help codify it. And then I think there also needs to be a bit of a conversation around, like, what qualifies as evidence and how CSOs can contribute to that. Because in the sort of for -profit AI sense, there's much more capital deployed to that than social initiatives. And if you were to, like, assume that ratio in terms of evidence, the evidence would be lacking. So I think, like, also building up, like, repeatable infrastructure to facilitate collaborations so people have access to more evidence would also be really important.
Dr. Pedro dos Santos Maia
Thank you so much, Addie. Vasileios, please go ahead.
Mr. Vasileios Rovilos
Absolutely. I can only echo what Emaan and Addie actually just already highlighted. Access to evidence, collaboration, truly fantastic elements to really bring together. To your question, I wouldn't. Do you? In such a binary. kind of like binary kind of like elements, either infrastructure element or, you know, access to evidence more concretely, I would actually try to bring those together. I would reframe it as in like not either or, but and. Like it's a collective element that one way or another brings this sense of independence in the element that, you know, CSOs actually provide a lot of evidence integrity and the way that they actually analyze, you know, that evidence integrity. Within a third common taxonomy or language, if you will, that actually speaks to what essentially, you know, gives reach. If you actually have an overly fragmented way of actually discussing within CSOs, civil society organizations as an ecosystem and actually try to extrapolate that, there was regulators across the globe, you actually, you know, risk having a bubble tower situation where no one actually speaks. And fragmentation is not good for a global commodity such as AI. So discussing AI governance more broadly. like that's an approach that we actually need to set aside one way or another and of course all that together brings influence if you have that independence that leads to evidence integrity and you have the common source language denominator that also opens doors for you basically and you know that you actually have already the legitimacy to bring all those great ideas out there and one way or another utilize that evidence to bring some substantive inputs and when it comes to barriers I fully agree with you on the access to evidence element is a deeply problematic element but one way or another what truly stands out to me is that if we actually verse it around you know that very well meant you know spirit of collaboration then you know loads of like minded actors can really make an impact when it comes to elements that one way or another will to a very influential position. And quite importantly, like in a bit of a proof point here, you know, like quite a few companies out there, like I think that it's roughly like a bit like north of 70 % rely on AI tools. And only four to 5 % of those companies that heavily rely on artificial intelligence solutions know how to actually conduct AI governance. And basically they rely on those tools without having the mature infrastructure to really control it. And that's deeply problematic. And despite the problematic element, I feel that to your point on collaboration, if we actually, you know, are trying to be hopeful about AI governance, that element of collaboration is going to be the key denominator that actually brings us together in one way or another. We resolve the issue more to the essence, more substantively.
Dr. Pedro dos Santos Maia
Thank you so much, Vasileios. So we need collaboration. We have the states also participating in this process. We need evidence. And how can we make AI governance? in a way that it can mirror youth participation? How can we bring them into this conversation? Is it in the same way that we've been working with experts? Do we need something more specific to cater to the youth? How do you all see this? I'm going to start again with Emaan, please. T
Ms. Emaan Yasin Malik
he question is about youth participation. So, in terms of youth participation, there are a few different things that come into play. So, obviously, we have students who are working, for example, I give the example of Pakistan. When I work with top high schools, we have students who have access to visas, language models, they have access to online courses. But then when I work with schools from kind of government -funded schools and underprivileged areas, there's a question of even how do you explain AI to them in Urdu, right? So when you talk about AI in Urdu I was trying to do an exercise with the kids And I thought of I was like how will I say AI So I said okay Nakli akal Which means fake knowledge Like that's what it kind of Was translating to And the students were like oh that sounds kind of scary So I thought that a lot of the knowledge that students Did have around AI was Maybe like creating images on TikTok or Things like that but they did not really have Knowledge about AI governance A lot of students that I work with think that They really need to know how to code in order to Have a say in this as well So I think that there are a lot of resources available But students have to be told about these One of the great Organizations that I work with is called Netmission Asia so it's a Hong Kong Based firm and they have A Netmission Academy For students who can Kind of join in obviously it's in English but they learn a lot About AI governance through that we also have great resources such as the Asia -Pacific Youth Internet Governance Forum. So they have a lot of resources available for students. And in fact, they also submitted a document for the WSIS process. So students were involved in that. Other than in Pakistan, like what I find is most useful to work with universities. So I do have our firm has 20 MOUs with universities. So we are working with them to create, we have human rights cells in these universities. So we are hoping to create like an AI and rights based module in that sense. But maybe coming from a biased standpoint, but I do think that, you know, every country does have their national human rights institution. Very important to encourage your specific country's NHRI to have an AI and rights based program model or at least a focal person so that these things can, can be addressed at all levels from an independent body.
Dr. Pedro dos Santos Maia
Fascinating, thank you so much, now pass the word to Addie.
Ms. Addie Achan
Thank you uh yeah thank you and it's really cool hearing how um it sort of is enacted in practice I think like what's really interesting about the youth perspective with AI is that compared to like most other industries we talk about I think like the functional difference between someone who's like 25 years old and someone who's 50 years old when it comes to using AI may not be that stark because it's like a relatively new technology um so I think it's a very important perspective to have especially when you think of more long -term effects like the environmental effects and whatnot like those particularly affect the youth um the question made me think of something that I researched before this panel which uh was a UNICEF essay written by Robert hart about children's participation and he basically broke down how youth can participate into eight rungs and at the bottom level is stuff like tokenism which he calls uh you apparently giving children the voice um with very little choice about subject style or opportunity to form an opinion but then on the other end of the spectrum it's just like having young people initiate and share decisions um so I think that's like a really powerful frame with AI I think um essentially like giving authority through various vehicles in the government um to allow youth to participate and I think like to make it really meaningful just like understand like what decision are they particularly changing and what are they informing what would they what would be lacking if they weren't there versus like would this be the same outcome without them here and is it performative and I think there is like probably a lot of work and even just like our internal biases around like how experience matters which like it does matter but I think experience is also comes from that younger age perspective.
Dr. Pedro dos Santos Maia
Fascinating, now please Vasileios.
Mr. Vasileios Rovilos
I couldn't I couldn't agree more and quite frankly uh I see a lot of distance in terms of like you know the decision makers uh and the youth element which is so concrete for what is actually coming our way uh I feel the But, you know, speaking to tokenization, that is like a very, you know, cheap way to actually occupy the frame as in like get people involved, but at the same time don't give them any sort of like concrete power or decision -making. And I feel that, you know, scaling the youth in various configurations that one way or another potentially have also this like fictional element is of true added value, but yet again we need to follow through. We need to actually get them involved into like processes that absolutely matter. And one way or another, that distance is structural. So I would love to actually just like put it out there. One way or another, it's important to be addressed. Now, I fully agree and resonate a lot with the point that Emaan was actually raising about, you know, like channeling resources, meaningful resources to the youth. Make sure that they're engaged. Make sure that one way or another there is this follow -up element that one way or another brings. Brings them into the picture, gets them engaged, gets them influential. And quite importantly, it's supposed to be an anchor or like... youth participation is supposed to be anchored into locally deployed AI systems. The case study of Pakistan is actually brilliant in many ways, and I was quite agnostic about it. I do have a really good overview of what happens in the EU with AI literacy campaigns and more concrete obligations within the EU AI Act and other regulatory frameworks. But yet again, it's quite important to actually make sure that we close the feedback loop with the people that are closest to those potentially even harmful AI case studies or tools. And versing the discussion back to the youth, I think it's very, very, very important to actually give them timing in the sense that we actually loop in the youth in the process proactively and preemptively. if everything is actually predetermined and one way or another simply need to live up to the expectations that others set for them and yet again they don't have any say into the decision making process that's going to get them even further away from the bull so that's definitely something that we actually want to avoid speaking to timing another element that really resonates with me is the influence part you give them influence not only by the decision making element but also making them aware of the fact that youth is the portion of the population that is most widely subjected to AI tools like you see that in proctoring tools for their homeschool activities you see that in any sort of algorithmic tool that they actually weaponize in their phones you see that down the line potentially because they're going to be progressing from being students to going to the actual markets and they're going to be hiring algorithmic tools that one way or another sort of affects their chances in the market and that's also quite important to mention here like making them aware of how important it is for them to claim that space, that decision -making space, is of utmost importance. And again, follow through. Having that follow -up, making sure that you actually close the feedback loop with the populations that are most affected, it's, you know, like it's a rather simple, you know, diagnostic. But it needs to be put out there. Quite a few actions out there. It's like, do you, like, walk the walk, but the... No, the talk the talk, but they do not walk the walk in a way. And that is kind of, like, disappointed. And in many levels, I would love to actually see the narrative being, like, completely, you know, like, it doesn't have space in AI governance more broadly. So I would love to actually address that as well. And from a European perspective, there are quite a few really cool activities I would love to actually urge you or to take a moment to actually look through. There's the AI Pact, which is a European Commission sort of activity that actually brings all the major, you know, technology actors and one way or another. I'd love to pay... the space, artificial intelligence or rely on such technologies to actually get together, have this multi -stakeholder group, bring in CSOs as well and youth to make sure that all together they co -design an element that actually speaks to their own needs. Because AI literacy done now is purely done for like deployers and very corporate actors. And, you know, youth again very much ostracized from the whole process. And that's to witness.
Dr. Pedro dos Santos Maia
Well, now I have one final question before the Q &A. And it's more of like an exercise of sorts. I want you three to imagine five years from now what do you want to see being implemented in terms of youth participation in AI governance? What types of bodies, what types of frameworks or benchmarks, what do you want to see being implemented in five years from now? I'll start with Emaan.
Ms. Emaan Yasin Malik
So, don't really have my thoughts gathered on this, but one thing I was thinking of throughout this process when I was hearing everybody speak. is that I really think that it has to be mandatory that there has to be youth participation in international reporting processes. And I think that, for example, for the UNCRC, we do include children in the consultation. But for digital -related international reporting processes, I think youth voice has to be made mandatory by the UN because then I think this pushes the state to actually provide evidence in terms of what the youth are saying. And I think that this should be in all reporting cycles, like in CAT, CERD, all of them. But, yes, this is the first thing I'm thinking of. In five years, maybe we can have this. Other than that, in five years, I would love to see an organization or maybe, I would say, maybe a commission run by the youth, a digital commission in Pakistan that can be built entirely. It should be independent, self -governed. They should be provided state resources to work on this matter because I think it's becoming really concerning. especially in a country like Pakistan that ranks really low in freedom of speech indexes where there's a lot of violence online, tech facilitated violence against children, minority communities, women I think this is really necessary so those are kind of the two things that I would like to see.
Dr. Pedro dos Santos Maia
Thank you so much Now I'm going to pass it over to Addie
Ms. Addie Achan
And maybe it's along specific issue areas that are relevant to them because you can't have a single thing that's going to encapsulate everything. But have, like, some type of, like, transparent participatory voting system where people can input what they would like to see from, you know, their various AI tools that are starting to enter their life. And then get very transparent objective scores around how well it's actually meeting those needs in a way that isn't biased by, like, AI companies themselves. So by an independent evaluation. That's what I hope to see in five years.
Dr. Pedro dos Santos Maia
Amazing. Now, Vasileios, please, go ahead.
Mr. Vasileios Rovilos
Yeah, sure. Absolutely. I fully concur with both of you. Like, it's a truly intergenerational issue after all, right? Like, if you think that through cutting youth from AI governance is the most sort of cited thing to ever even think of. And one way or another, you know, we started the conversation with bringing in the picture that the main barrier is access to evidence. But yet again, funds. We could actually invest more funds. in that direction. So that's a concrete element that we'd love to actually see not being downplayed in the coming five years. Again, viewing how the EU actually does things, but also the EU 27 member states across the continent view these talents. They tend to actually have, again, the Roger Hurti report elements, like the tokenization element. It's actually putting people up there to actually bring their own concerns but yet again, it actually ends up being very performative because there's no follow -through. I would love to see that follow -through. And in that sense, I fully agree with both of your suggestions to actually have that impact -driven approach in including youth into the decision -making process. And even with accentuated or even augmented rights to actually do that. It would be fantastic to actually see. And one way or another, I feel that ... It all comes back to also AI literacy in a way because, you know, we are so preoccupied about, like, productivity gains within our corporate kind of like lives. But, again, that's society. Like the people, the intergenerational element comes in play again. Youth is the one portion of the population. And the youth after them, that one way or another will need to figure out a forward -leaning way to actually address, you know, all potential implications and impact that comes with relying on artificial intelligence tools. And one way or another, skilling the youth would be of utmost importance. So I would love to actually see that be more, you know, into the day, like ingrained into the day -to -day in the way that the institutions thinks, the way that they actually act, and the way that they actually decide. So...
Dr. Pedro dos Santos Maia
So fascinating stuff, a lot of food for thought. We have 10 minutes for the Q &A, but before giving the floor to the audience, I'd like to ask for a round of applause for our panelists. Thank you so much. So now to the audience, does anyone have questions? We can start with you.
Audience
Thank you. Thank you all. I have a question that . You know that the co -founder and the CEO of Google, he believes that maybe will come before 2030. So if will come in five years, what will that will be and how can we do AI governance? Are young people doing that?
Dr. Pedro dos Santos Maia
I'm going to get another question, then I'm going to pass the words to our panelists. Who has questions?
Audience
So, my question is, you talked a lot about the intergenerational dialogue that needs to take place, and I totally agree with that. But how do we foster that dialogue? And are teachers and parents equipped to actually teach AI to the next generation? And if not, how do we kind of upscale that?
Dr. Pedro dos Santos Maia
Thank you so much. Now, please. Who wants to go first?
Ms. Addie Achan
Okay, I can jump in. So, I mean, I think, like, it's interesting. I don't know, but, like, in a way, I think it addresses both, maybe not, like, fully directly head -on, but I think, like, to answer your question directly, I don't think teachers and parents are equipped. It's like a new skill set, right? So there are some people that really understand this skill set, and I think it has very little to do with age, given how recent the technology is. Which, maybe it doesn't provide an exact solution around like who's the one arbitrating it, but I think like it gives more credibility to the idea that that needs to be given an equal voice and all those need that participation. I think when it comes to AGI and governance, I think like even right now we're kind of like in a moment where knowledge is like getting commoditized, like there's obviously a lot of value in like, you know, knowing your own stuff. And I would like to think that that discernment piece is always going to be important. Obviously, it's really difficult to understand like exactly what that looks like in practice. But I'm like, even if the models themselves are capable of doing more, I think there are still like so many perspectives about how that affects the human experience that need to be governed. And I think youth have a very important role in that as again, it'll be like they're understanding it at the same time as everyone else, perhaps with like a bit more like flexibility. in their mind um but I also think that will require uh that that intergenerational collaboration so i think like youth will have an important role I don't think it's an exclusive role.
Dr. Pedro dos Santos Maia
Anyone else please.
Ms. Emaan Yasin Malik
Sure so um I can give an example for your question about parents and teachers so I think that teachers are overworked they're the pillars of our society and they welcome a tool that will help them and they were that will assist them I work with a bunch of schools at one specific high school in Pakistan lahore grammar school we are having admissions these days so all of the parents that are coming in they're like how do we get our kid into AI like how can you help our kid get into AI and and it is it's a fear in the parents right because they don't necessarily have the language to um kind of describe what this phenomenon is but they're seeing it on tv in the news there's conflicting new news in the newspapers there's a new data center being built in Pakistan there's water shortage they're like what's happening our kid needs to be ahead so I think that there's a willingness there's excitement and I think both among students and teachers. So I think that if we do provide them resources, if we provide them curriculum, if we provide them maybe designs for workshops, if we make these things really easily accessible and free, I think that teachers are really willing to implement these in schools. And of course, if we make these language friendly, I think that there's... And I think this is an ITU stat. I think they had a really interesting survey which was ranking excitement about AI, and it was really high in South Asia as compared to fears. I think there's scope and positivity towards this question.
Mr. Vasileios Rovilos
Yeah, I'd love to actually address both questions on the AGI elements. Quite a few elements to actually consider here. Okay, there's the projection that maybe in five years this is going to come into play, and it actually comes from a very, very credible source as well. So like, yeah, sure. But yet again, I feel that we're just like, In terms of setting societal expectations, making sure that we also capture the picture holistically and properly in terms of regulations on all things artificial intelligence is of utmost importance. We barely managed in the EU to actually capture the EU AI Act, which actually puts a classified risk. And based on the classification that you get, you actually get a set of obligations that you absolutely need to live up to. And of course, there are prohibited practices as well. And what comes in mind, for me at least, on the AGI elements is that AGI potentially will completely disregard the element of human oversight. And that does not necessarily resonate with the way the AI tools will be marketed. broadly as a global commodity. So, like, many, many, many elements to actually think about it. And quite frankly, like, we have few pertinent challenges. Like, potentially you actually saw what happened in the U .S. with the export controls on Mythos and 505, and that is widely geopolitically driven. But yet again, it actually illustrates the institutional muscle that comes with governments withholding such tools from being marketed or put in service within X market globally. So there are quite a few. Like, so we have a lot of ground to cover, a lot of space to cover. And my conservative approach would be step by step and one step ahead of each other, like in front of the other, instead of, like, leaping ahead with, like, capturing all these, like, looming elements. Nonetheless, what I feel that is super important to actually, well, we should actually get our ducks in a row. is the element where we actually set common expectations. Emaan before actually spoke to what actually works, you know, in AI and AI governance more broadly. And if we don't have that very wide multi -stakeholderism approach in reading that situation, then we're actually led to like a wall and an impasse. And one way or another, that's also a very important consideration to take into account. And that's a spirit initiative, right? Yeah, I do remember you. I don't know why, but I do remember you. So brilliant question. And one thing that actually comes to mind instantly, I was actually back home in Greece and I was just visiting my family. I saw my niece and she was with your tablet actually doing, she was playing with cloud, code. and I asked her what do you do because I actually just like try to do that for my work too one way or another and I asked her like who taught you to do that and she's like I asked Claude and I'm like so you have no direct supervision from your school teachers because she was actually weaponizing the tool for homework smart right she tells me they don't know and I'm like that is the answer to your question like one way or another I feel that investing heavily on making sure that teachers, educators one way or another are within the remit of people that need to live up to the expectations of AI literacy is like my top priority I would say in terms of AI literacy and quite importantly AI literacy also covers as an obligation or should be in my opinion covering as an obligation public or bodies that are dictated by public law so schools should be a top tier priority for that element.
Dr. Pedro dos Santos Maia
Well, due to time constraints, I want to finish on time. And please do not hesitate to approach us, to talk to our panelists. We will be hanging here after the panel. Thank you so much for coming. I would like to thank as well Globefix for putting all this together. And before we leave, one final round of applause for our panelists, please. Thank you. Thank you all for coming. Please, again, we are all here. See you around. Take care. Have a good day. Bye -bye.
Inclusive AI governance: Universal values in a pluralistic world
These values can inform governance models that prioritise relational accountability, ethical cultivation, and social cohesion, offering alternatives to transactional, compliance-driven frameworks.This is why I dare here ...
Navigating the interplay between artificial intelligence, philosophy, education, and governance
Overview The rapid development of artificial intelligence (AI) presents profound philosophical, educational, and governance challenges and opportunities for the global community. To address these dimensions, we organis...
Global AI governance: Reflecting on 2024 and shaping the path for 2025
On 9 January 2025, DiploFoundation, in collaboration with the Permanent Missions of China, France, Kenya, Mexico, Pakistan, Switzerland, and the United States to the United Nations in Geneva – as co-sponsors – hosted an ...
Main Session 2: The governance of artificial intelligence
Main Session 2 Rapport de session Intervenants Graphe de connaissances Analyse approfondie Aperçu de la session Résumé This discussion fo...
AI promises, ethics, and human rights: Time to open Pandora’s box
Awareness and training As a sociotechnical system, AI depends on goals, data sets, and contexts in which it is deployed. Impacts (positive or negative) are a reflection of the designers' and operators' values. Diversit...
The MANAV manifesto: Reclaiming agency for the majority
India AI Impact Summit 2026 offered something rare in the global AI conversation: a blueprint that begins not with the algorithm, but with the human. The word manav means 'human' or 'mankind' in Sanskrit. It is a deli...
Review of AI and digital developments in 2024
They are becoming de facto content regulators, determining what content is allowed and what is removed from their platforms. In 2024, these companies are expected to continue adapting their content moderation policies to...
WS #236 Ensuring Human Rights and Inclusion: An Algorithmic Strategy
And one of the most challenging things is AI policy. And I do believe that protecting democracy, human rights, the rule of law, and establishing clear guidelines on AI is a shared responsibility that a government alon...
Keeping AI in check
A ten-step guide published by CoE starts with the need to conduct a human rights impact assessment on AI systems. Technology is a social product, and as such, it integrates values that orientate the way it operates, ev...
Four seasons of AI:  From excitement to clarity in the first year of ChatGPT
How to address AI risks There are three main types of AI risks that should shape AI regulations: the immediate and short-term ‘known knowns’ the looming and mid-term ‘known unknowns’ and the long-term yet dis...
WS #187 Bridging Internet AI Governance From Theory to Practice
This approach would provide transparency about safety measures without exposing the entire model architecture. Preuves Suggests requiring AI developers to implement robust safety guardrails and publish information a...
Human rights
Clear frameworks for accountability and oversight are necessary to address issues arising from AI's use. 5. Legal and Regulatory Frameworks: Guidelines and Regulations: Strong rules and guidelines are needed for develop...
WS #438 Digital Dilemmaai Ethical Foresight Vs Regulatory Roulette
She sees this as a foundational requirement before other governance measures can be effective. Major discussion point Infrastructure and Capacity Building Topics Infrastructure | Development MModeratorSpeech spe...
WS #266 Empowering Civil Society: Bridging Gaps in Policy Influence
Language barriers and technical jargon: Kenneth Harry Msiska noted that complex technical language often alienates CSOs, making it difficult for them to engage effectively in discussions. Lack of technical capacity: ...
How to enhance participation and cooperation of CSOs in/with multistakeholder IG forums | IGF 2023 Open Forum #96
Consultation processes prior to the introduction of a bill in India, for example, have proven to be more fruitful in generating meaningful engagement.Finding ways to engage directly with parliamentarians and government o...
Diplo at IGF 2024
If so, how much are they reflected/considered in our AI policy and governance discussions? Assigning human attributes to AI: What do we talk about when we talk about AI ‘understanding’, ‘reasoning’, etc.? When words lose...
Meaningful Youth Engagement in Policy and Decision-making Processes | Our Common Agenda Policy Brief 3
The Global Coalition on Youth, Peace and Security plays a key role in facilitating this work. Elsewhere, however, youth participation is less structured. For instance, there is no formal mechanism for youth to engage s...
Multilateral Intergenerational High-Level Dialogue: Youth Special Track
Preuves Founder of Youth Climate Voice Caribbean, leading digital campaigns and webinars that connect young people with climate data, early warning systems, and SDG storytelling across the Caribbean region Major dis...
1

The knowledge base provides broader support for this concern by noting that simply transferring technologies from developed countries to developing countries is often inappropriate because such transfers require context-specific considerations [S85]. It also warns that developing states risk becoming providers of raw data while paying for intelligence services developed elsewhere [S87].

2

This is consistent with wider AI governance discussions in the knowledge base stressing the need for practical tools, resource libraries, standards, and reliable assessment methods for testing AI systems and translating principles into implementation [S89].

3

The emphasis on actionable mechanisms aligns with broader knowledge-base discussions that AI governance now needs to move from high-level principles to practical, implementable measures and toolkits [S89].

4

The knowledge base supports the general pattern that women and minorities are often under-represented in technology sectors and decision-making, and that diversity in AI teams matters for inclusion and rights-sensitive outcomes [S97] and [S39].

European Broadcasting Union — Capacity development Most of the EBU’s activities are aimed at increasing the capacity of its members to address challenges and embrace opportunities brought about by the digital age. To that end, through its Digital T...
EQUAL Global Partnership Research Coalition Annual Meeting | IGF 2023 — Is there any summary from Audience: their discussions? Yeah maybe I can just briefly summarize so thanks to the great tech team we were able to like solve our issues and communicate with each other but...
Vitali Francesco — Vitali Francesco
Ilias Dimou — https://diplo-media.s3.eu-central-1.amazonaws.com/2024/09/Ilias-Dimou.jpg Mr Ilias Dimou is a political scientist with a strong academic background, holding master's degrees in Ethics and Political Philosophy, as well as...
Luka Avramović — https://diplo-media.s3.eu-central-1.amazonaws.com/2024/04/Luka-Avramovic.png Mr Luka Avramović is currently undertaking an MA in Politics at the University of Glasgow. Previously, he held positions as a communications as...
Pedro Vilela Resende Gonçalves — Pedro Vilela Resende Gonçalves https://www.diplomacy.edu/wp-content/uploads/2018/09/Pedro-Vilela-Resende-Gonçalves_small_square.png
Pedro Gomez — Pedro Gomez
Communications and competition law: Key issues in the telecoms, media and technology sectors — Mr. Moura holds a BA degree in Economics and an MBA from COPPEAD – Rio de Janeiro Federal University. Prior to joining TelComp, Mr. Moura was a Partner at Coopers & Lybrand – a global consultancy firm – with significant ...
Yasmin Ismail — Ms Yasmin Ismail holds two Master’s degrees in Economics of Development and International and European Law from the University of Grenoble-Alpes in France. Yasmin was a Program Officer at CUTS International, Geneva, and ...
Kaarika Das — https://diplo-media.s3.eu-central-1.amazonaws.com/2023/09/Kaarika-Das-1.jpg Ms Kaarika Das is a PhD candidate in Economics of Education at the National Institute of Education Planning and Administration in New Delhi. Her...
World Economic Forum – Global Coalition for Digital Safety | IGF 2023 Side Event — https://www.intgovforum.org/en/content/enhancing-digital-safety-the-world-economic-forum-global-coalitions-collaborative-approach Intervenants: Name: Courtney Gregoire, Affiliation: Chief Digital Safety Officer, Microsof...
A Digital Future for All (morning sessions) — We all have a role to play in this global challenge. Hannah was able to rent an apartment where she now lives with her two children, her elderly mother, and her grandmother. We support people like Hannah who have been ...
Adda B. Bozeman — Adda B. Bozeman
Closing Ceremony and Orientation for WAIGF 2025 — Okay, I want to be part of IGF at national level. I want to be part of IGF in my country, part of IGF in my country, because we are not just here to do this for formality. We want to prepare you. You can see that I'm a g...
AI Accountability and Digital Trade: Policy Options for the UK — Everyday interactions are mediated by algorithms, with significant impacts for the individual, economies, and wider society. The AI-as-a-service global market is growing rapidly, and governments are struggling to regulat...
Workshop 1: AI & non-discrimination in digital spaces: from prevention to redress — Today we're here to explore a critical and urgent question. How can we ensure that AI systems, which are now increasingly used in law enforcement, in welfare and employment and in health, do not infringe or amplify discr...
Information Integrity on Digital Platforms | Our Common Agenda Policy Brief 8 — Available at https://ec.europa.eu/commission/presscorner/detail/en/mex_23_723. DIGITAL PLATFORM RESPONSES Digital platforms are very diverse in terms of size, function and structure and have pursued a wide range o...
WS #266 Empowering Civil Society: Bridging Gaps in Policy Influence — Civil society is the least positioned to hire people who have high technical expertise. Look at the money that big tech is paying and compared to the money that civil society organizations would pay. I mean in my c...
Preuves and measurement in Internet governance — In every policy area, informed policy-making requires access to data and evidence. Solutions simply cannot be advanced without an accurate scope of the problem. Internet governance and ICT policy are no exception. For po...
The State of the Humanitarian System — in relation to flash appeals). This was manifested in a greater percentage of total annual contributions being Evaluations tend to focus on the strengths and committed in the first and second quarters ...
WS #65 Gender Prioritization through Responsible Digital Governance — Which means that there is an urge, there is a requirement by the women but they don't have the phone so they make use of the phones that the family has and they use it. But the good thing was those females who have the...
Addressing the Digital Divide in the Joint Statement Initiative on E-Commerce: from enabling issues to data and source code provisions — 2.6 E-commerce and Closing the Gender Gap in Trade: Toward gender-inclusive e-commerce As mentioned above, new business models are emerging due to the fast digitalization of the economy and the rise of e-commerce. Some o...
Main Session | Policy Network on Meaningful Access — That is director of the Pakistan Telecom Authority in charge of wireless. And the experience that you will present us concerns gender, because of course, there is a problem with gender. We have seen yesterday the st...
Youth and media literacy: EU – US lessons and practices — In the World Summit on the Information Society (WSIS) Tunis Agenda, children were not mentioned as a stakeholder group, though referred to as a vulnerable group needing protection. Nowadays, global governance initiative...
AI for Good Global Summit — And then we need to be the silence change agent, right? Maybe we have to say it, but do it more, right? I think that's gonna be an interesting time. We'll repeat it as in Haiti, Haitian Creole, which is not a standard la...
Meaningful Youth Engagement in Policy and Decision-making Processes | Our Common Agenda Policy Brief 3 — Kigali, 2022) of the World Telecommunication Development Conference.PROGRAMME OF ACTIONOF THE INTERNATIONALCONFERENCE ONPOPULATION ANDDEVELOPMENTHigh-Level Commissionon the NairobiSummitProvides high-level leadership and...
Rethinking learning: Hope, solutions, and wisdom with AI in the classroom — My friend's experience suggests the answer: everything that makes education actually work. She recognises when students have used AI not through sophisticated detection software but through human observation, noticing ...
AI in schools: The reality is messier than the solutions — As the school year is in full swing, the issue of AI in schools and education keeps coming up everywhere. Teachers share stories in faculty lounges, parents worry at dinner tables, and students find themselves in a chall...
Deepfakes and the AI scam wave eroding trust — Doing so requires more than better detection tools. It calls for technical systems that give genuine content a clear signal of authenticity, ethical norms that discourage irresponsible use of generative AI, and instituti...
Local, Everywhere: The blueprint for a Humanitarian AI transformation — The slogan'Local, everywhere' of the 2026 Plan of the International Federation of Red Cross and Red Crescent Societies (IFRC) captures the essential blueprint for the bottom-up development of Humanitarian AI: technology ...
Towards 2030 and Beyond: Accelerating the SDGs through Access to Preuves on What Works — And I thank you again very much for the opportunity of taking part of the conversation. Andrea Cook: Thank you, Ambassador Jiménez de La hoz, and for grounding your comments in the actual experience that we've been ...
AI, smart cities, and the surveillance trade-off — The algorithm identifies patterns and extrapolates them into the future. This sounds rational until you consider why those patterns exist in the first place. Neighbourhoods that were historically neglected may show hig...
Keeping AI in check — A ten-step guide published by CoE starts with the need to conduct a human rights impact assessment on AI systems. Technology is a social product, and as such, it integrates values that orientate the way it operates, ev...
Inclusive AI governance: Universal values in a pluralistic world — These values can inform governance models that prioritise relational accountability, ethical cultivation, and social cohesion, offering alternatives to transactional, compliance-driven frameworks.This is why I dare here ...
Navigating the interplay between artificial intelligence, philosophy, education, and governance — Overview The rapid development of artificial intelligence (AI) presents profound philosophical, educational, and governance challenges and opportunities for the global community. To address these dimensions, we organis...
Global AI governance: Reflecting on 2024 and shaping the path for 2025 — On 9 January 2025, DiploFoundation, in collaboration with the Permanent Missions of China, France, Kenya, Mexico, Pakistan, Switzerland, and the United States to the United Nations in Geneva – as co-sponsors – hosted an ...
AI and international peace and security: Key issues and relevance for Geneva — This includes education, public-private partnerships, regional cooperation, and international dialogues. By advancing these efforts collaboratively, states can ensure that all stakeholders are equipped to participate mea...
AI promises, ethics, and human rights: Time to open Pandora’s box — Awareness and training As a sociotechnical system, AI depends on goals, data sets, and contexts in which it is deployed. Impacts (positive or negative) are a reflection of the designers' and operators' values. Diversit...
Day 0 Event #174 Human Rights Impacts of AI on Marginalized Populations — This includes issues like deep fake images and videos that disproportionately affect women and girls. Preuves Experience from META's oversight board in framing how companies like Meta can use automation to deal with...
The MANAV manifesto: Reclaiming agency for the majority — India AI Impact Summit 2026 offered something rare in the global AI conversation: a blueprint that begins not with the algorithm, but with the human. The word manav means 'human' or 'mankind' in Sanskrit. It is a deli...
Review of AI and digital developments in 2024 — They are becoming de facto content regulators, determining what content is allowed and what is removed from their platforms. In 2024, these companies are expected to continue adapting their content moderation policies to...
Four seasons of AI:  From excitement to clarity in the first year of ChatGPT — How to address AI risks There are three main types of AI risks that should shape AI regulations: the immediate and short-term ‘known knowns’ the looming and mid-term ‘known unknowns’ and the long-term yet dis...
Human rights — Clear frameworks for accountability and oversight are necessary to address issues arising from AI's use. 5. Legal and Regulatory Frameworks: Guidelines and Regulations: Strong rules and guidelines are needed for develop...
WS #187 Bridging Internet AI Governance From Theory to Practice — This approach would provide transparency about safety measures without exposing the entire model architecture. Preuves Suggests requiring AI developers to implement robust safety guardrails and publish information a...
How to enhance participation and cooperation of CSOs in/with multistakeholder IG forums | IGF 2023 Open Forum #96 — Consultation processes prior to the introduction of a bill in India, for example, have proven to be more fruitful in generating meaningful engagement.Finding ways to engage directly with parliamentarians and government o...
Diplo at IGF 2024 — If so, how much are they reflected/considered in our AI policy and governance discussions? Assigning human attributes to AI: What do we talk about when we talk about AI ‘understanding’, ‘reasoning’, etc.? When words lose...
Multilateral Intergenerational High-Level Dialogue: Youth Special Track — Preuves Founder of Youth Climate Voice Caribbean, leading digital campaigns and webinars that connect young people with climate data, early warning systems, and SDG storytelling across the Caribbean region Major dis...
Inclusive AI governance: Universal values in a pluralistic world — These values can inform governance models that prioritise relational accountability, ethical cultivation, and social cohesion, offering alternatives to transactional, compliance-driven frameworks.This is why I dare here ...
Human rights — Clear frameworks for accountability and oversight are necessary to address issues arising from AI's use. 5. Legal and Regulatory Frameworks: Guidelines and Regulations: Strong rules and guidelines are needed for develop...
Four seasons of AI:  From excitement to clarity in the first year of ChatGPT — How to address AI risks There are three main types of AI risks that should shape AI regulations: the immediate and short-term ‘known knowns’ the looming and mid-term ‘known unknowns’ and the long-term yet dis...
AI diplomacy — Privacy and data protection are particularly pertinent, given that AI systems often require massive datasets, which can include sensitive personal information. It is crucial that developers embed privacy guarantees and d...
Meaningful Youth Engagement in Policy and Decision-making Processes | Our Common Agenda Policy Brief 3 — A significant proportion of Governments have put in place national youth policiesSee A/70/156 and many have also established mechanisms to facilitate more structured participation of young people in decision-making (see ...
Multilateral Intergenerational High-Level Dialogue: Youth Special Track — Preuves Founder of Youth Climate Voice Caribbean, leading digital campaigns and webinars that connect young people with climate data, early warning systems, and SDG storytelling across the Caribbean region Major dis...
Rethinking learning: Hope, solutions, and wisdom with AI in the classroom — My friend's experience suggests the answer: everything that makes education actually work. She recognises when students have used AI not through sophisticated detection software but through human observation, noticing ...
AI in schools: The reality is messier than the solutions — As the school year is in full swing, the issue of AI in schools and education keeps coming up everywhere. Teachers share stories in faculty lounges, parents worry at dinner tables, and students find themselves in a chall...
Beyond partnership and participation: Community organising for ownership in humanitarian diplomacy and aid — A potentially highly relevant approach in an international context is Slum Dwellers International (SDI). SDI defines itself as ‘a network of community-based organisations of the urban poor in 32 countries and hundreds of...
Morning session — which could potentially constrain their autonomy. This suggests genuine recognition of the value of inclusive governance models Topics Legal and regulatory | Economic Civil society and government alignment on digital ...
Nri Collaborative Session Community Based Connectivity to Achieve Digitally Inclusive Societies — This represents a shift from traditional views where civil society responds to government policies to one where they actively shape them Topics Legal and regulatory | Development | Sociocultural Community networks are...
AI, smart cities, and the surveillance trade-off — The algorithm identifies patterns and extrapolates them into the future. This sounds rational until you consider why those patterns exist in the first place. Neighbourhoods that were historically neglected may show hig...
The MANAV manifesto: Reclaiming agency for the majority — There is a phrase that captures this orientation well: the 'right to be imperfect'. It is the idea that AI systems must make space for the full spectrum of human users, not just those who are highly literate, highly conn...
The year of AI clarity: 10 AI Forecasts for 2025 — China: China has implemented strict regulations requiring platforms to label AI-generated content, especially deepfakes, and to obtain consent from individuals before using their likenesses. Which practices do social ...
Addressing AI-driven Socio-economic Inequality — In March 2020, entrepreneur and billionaire Elon Musk stated that 'AI is far more dangerous than nukes'. This wasn't he first time the adverse effects of the scale of AI deployment has been discussed. Throughout the last...
IGF 2024 Global Youth Summit — Preuves The speaker mentions specific roles for developers (designing ethical and unbiased systems), educators (implementing AI effectively), policymakers (creating regulations and guidelines), and students (understan...
Digital literacy for digital natives — Since its creation, the Internet has been regarded as a world of opportunities. Whereas it gives access to information, knowledge, and ideas, and contributes to faster communication and new business models, online engage...
Inclusive AI governance: Universal values in a pluralistic world — These values can inform governance models that prioritise relational accountability, ethical cultivation, and social cohesion, offering alternatives to transactional, compliance-driven frameworks.This is why I dare here ...
Navigating the interplay between artificial intelligence, philosophy, education, and governance — Overview The rapid development of artificial intelligence (AI) presents profound philosophical, educational, and governance challenges and opportunities for the global community. To address these dimensions, we organis...
Global AI governance: Reflecting on 2024 and shaping the path for 2025 — On 9 January 2025, DiploFoundation, in collaboration with the Permanent Missions of China, France, Kenya, Mexico, Pakistan, Switzerland, and the United States to the United Nations in Geneva – as co-sponsors – hosted an ...
Main Session 2: The governance of artificial intelligence Main Session 2 Rapport de session Intervenants Graphe de connaissances Analyse approfondie Aperçu de la session Résumé This discussion fo...
AI promises, ethics, and human rights: Time to open Pandora’s box — Awareness and training As a sociotechnical system, AI depends on goals, data sets, and contexts in which it is deployed. Impacts (positive or negative) are a reflection of the designers' and operators' values. Diversit...
The MANAV manifesto: Reclaiming agency for the majority — India AI Impact Summit 2026 offered something rare in the global AI conversation: a blueprint that begins not with the algorithm, but with the human. The word manav means 'human' or 'mankind' in Sanskrit. It is a deli...
Review of AI and digital developments in 2024 — They are becoming de facto content regulators, determining what content is allowed and what is removed from their platforms. In 2024, these companies are expected to continue adapting their content moderation policies to...
WS #236 Ensuring Human Rights and Inclusion: An Algorithmic Strategy — And one of the most challenging things is AI policy. And I do believe that protecting democracy, human rights, the rule of law, and establishing clear guidelines on AI is a shared responsibility that a government alon...
Keeping AI in check — A ten-step guide published by CoE starts with the need to conduct a human rights impact assessment on AI systems. Technology is a social product, and as such, it integrates values that orientate the way it operates, ev...
Four seasons of AI:  From excitement to clarity in the first year of ChatGPT — How to address AI risks There are three main types of AI risks that should shape AI regulations: the immediate and short-term ‘known knowns’ the looming and mid-term ‘known unknowns’ and the long-term yet dis...
WS #187 Bridging Internet AI Governance From Theory to Practice — This approach would provide transparency about safety measures without exposing the entire model architecture. Preuves Suggests requiring AI developers to implement robust safety guardrails and publish information a...
Human rights — Clear frameworks for accountability and oversight are necessary to address issues arising from AI's use. 5. Legal and Regulatory Frameworks: Guidelines and Regulations: Strong rules and guidelines are needed for develop...
WS #438 Digital Dilemmaai Ethical Foresight Vs Regulatory Roulette — She sees this as a foundational requirement before other governance measures can be effective. Major discussion point Infrastructure and Capacity Building Topics Infrastructure | Development MModeratorSpeech spe...
WS #266 Empowering Civil Society: Bridging Gaps in Policy Influence — Language barriers and technical jargon: Kenneth Harry Msiska noted that complex technical language often alienates CSOs, making it difficult for them to engage effectively in discussions. Lack of technical capacity: ...
How to enhance participation and cooperation of CSOs in/with multistakeholder IG forums | IGF 2023 Open Forum #96 — Consultation processes prior to the introduction of a bill in India, for example, have proven to be more fruitful in generating meaningful engagement.Finding ways to engage directly with parliamentarians and government o...
Diplo at IGF 2024 — If so, how much are they reflected/considered in our AI policy and governance discussions? Assigning human attributes to AI: What do we talk about when we talk about AI ‘understanding’, ‘reasoning’, etc.? When words lose...
Meaningful Youth Engagement in Policy and Decision-making Processes | Our Common Agenda Policy Brief 3 — The Global Coalition on Youth, Peace and Security plays a key role in facilitating this work. Elsewhere, however, youth participation is less structured. For instance, there is no formal mechanism for youth to engage s...
Multilateral Intergenerational High-Level Dialogue: Youth Special Track — Preuves Founder of Youth Climate Voice Caribbean, leading digital campaigns and webinars that connect young people with climate data, early warning systems, and SDG storytelling across the Caribbean region Major dis...
Main Session on Artificial Intelligence | IGF 2023 — I'm very pleased to be here and have the pleasure of being the moderator of this session, which is about the digital transformation, and the digital transformation that we want. So I have the honour to have a distinguish...
5th 'Road to Bern via Geneva' dialogue: On data and Tech4Good — She stated that efforts should be made to bridge these gaps. For achieving the SDGs, the ‘business-as-usual’ approach and relying on existing solutions is not sufficient. Simply delivering technologies available in devel...
Regional experiences on the governance of emerging technologies NRI Collaborative Session — There is no model, I think that's the. message. There is guidelines, but there is no model that fits all. That's it. Thank you. Ano Jawhar Abdulmaseeh Abdoka: To sum up, the magical word is collaboration and mutual ...
Data Governance in the Context of Emerging Technologies: Promoting Human-Centred and Development-Oriented Societies   — He also mentioned that finding a single definition is challenging as data can be looked at from different angles. For example, the way data is seen from an economic perspective (which is the focus of the course) is diffe...
[WebDebate #26 summary] AI on the international agenda - where do we go from here? — In our December WebDebate, we looked at artificial intelligence (AI). AI is of geostrategic importance and countries are already investing heavily in AI and developing AI strategies. As AI potentially impacts nearly all ...
Open Forum #30 High Level Review of AI Governance Including the Discussion — with global initiatives providing both direct guidance and indirect harmonization across jurisdictions Résumé All intervenants agree that while AI governance principles have been established across various frameworks, the ...
General Elections in Pakistan: An Indian Perspective — The world attention was focussed on the historic General Elections in Pakistan in May 2013, amidst an air of excitement and trepidation, as these signified transfer of civilian power, for the first time in the 67 years ...
Diplomatic realism: Nixon, Kissinger, and Pakistan — Kissinger demanded that Nixon stand firm [in supporting Pakistan]. 'If the outcome of this war is that Pakistan is swallowed up by India, China is destroyed, defeated, humiliated by the Soviet Union, it will be a change ...
Opening of the session — Their positive outlook is evident in their endorsement of Article 5 as well as Article 24 within the draft of a Convention designed to safeguard human rights on a global scale.Regarding Article 5, the Nigerian representa...
The use of ICT in human rights promotion: A case study of the African Commission on Human and Peoples’ Rights — https://www.diplomacy.edu/wp-content/uploads/2021/06/IGCBP2010_2011_Nsega.pdf https://www.diplomacy.edu/wp-content/uploads/2021/06/page_1-4.jpg
The use of Information and Communications Technology (ICTs) in human rights promotion: A case study of the African Commission on Human and Peoples' Rights — https://www.diplomacy.edu/wp-content/uploads/2021/06/11031016115009_Nsega_28Library29.pdf https://www.diplomacy.edu/wp-content/uploads/2021/06/umcrest.png
Inter-American Commission on Human Rights — The Inter-American Commission on Human Rights is an autonomous organ of the Organization of American States. The separate Inter-American Court of Human Rights is an autonomous judicial institution based in the city of Sa...
Impact of measures to address terrorism and violent extremism on civic space and the rights of civil society actors and human rights defenders — 18 A/HRC/40/52 system, except when requested by the assessed Member State to keep selected information confident...
‘Women still under-represented in the ICT sector’ - Kroes — European Commissioner Vice-President Neelie Kroes has blogged about it again: women are still under-represented in the ICT sector. I continue to find it troubling that women are under-represented in the ICT sector at ev...
UN Women and IamtheCODE hold hackathon in South Africa to help bridge gender digital divide — On 24–25 March 2023, the UN Women Regional Office for East and Southern Africa and IamtheCODE (a global movement dedicated to advancing science, technology, engineering, arts, mathematics and design education) held a two...
YouthLead: Inclusive digital future for all — And that notwithstanding, he created a Ministry of Communication and Innovation, appointed a young lady at the age of 27, to ensure that she leads that area. So, from the area of government, what we believe is that young...
WS #166 Breaking Barriers: Empowering Women in Internet Network — She introduced the Digital Egypt generations program and advocated for flexible work policies. 3. Arinola Akinyemi (Nigeria): Addressed workplace discrimination and harassment, proposing incentives for companies that e...
How does identity impact segregation? — 'Identity politics is back', opines a comment to my blog post The Trump Swerve. I admit to having difficulties with the concept of 'identity'. To put my queasiness into perspective, let me take a roundabout way and refle...
Workshop 5: Bridging Digital Inequalities and Challenges in Multicultural Societies — It forced intervenants to confront the possibility that despite numerous well-intentioned programs, systemic inequalities persist. This meta-question about the field's effectiveness was particularly thought-provoking. Impa...
Towards equitable and sustainable community-led networks — Citing her experiences in Kenya and in Siberia, she said that women are still economically dependent on men due to the wage gap, which affects the ability of women to afford smartphones and other devices that serve as a ...
The New Delhi AI Summit: Inclusive rhetoric, fractured reality — In the words of UN Secretary General, António Guterres, “If we want AI to serve humanity, policy cannot be built on guesswork. It cannot be built on hype – or disinformation. We need facts we can trust – and share – acro...
AI and the Roadmap for Digital Cooperation — Second, values that should guide the development and application of AI were touched upon in almost all interventions. Mr Max Tegmark (President, Future of Life Institute) expressed a key concern, shared by many, when he ...

Avertissement : Il ne s'agit pas d'un compte rendu officiel de la session. DiploAI génère ces ressources à partir d'enregistrements audiovisuels ; elles sont présentées telles quelles, y compris d'éventuelles erreurs. En raison de contraintes logistiques (audio/vidéo ou transcriptions), les noms peuvent être mal orthographiés. Nous nous efforçons d'être aussi précis que possible.