Leaders TalkX: Local to global: preserving culture and language in a digital era

Summary

This Leaders’ Talk focused on preserving cultural and linguistic diversity in the digital era, particularly as artificial intelligence becomes increasingly dominant in how people access information and services. The discussion highlighted that while this challenge began over 20 years ago with the rise of the internet, it has become more critical today with AI-driven technologies that are predominantly trained on English-language content, representing over 90% of large language models.

Gatis Ozols from Latvia described his country’s proactive approach to digitalizing the Latvian language, emphasizing that multilinguality must be viewed as a core principle rather than just a feature of AI systems. He outlined Latvia’s strategy of developing national language resources, supporting machine translation, and ensuring citizens can access government services in their native language. Pierre Bonis from AFNIC stressed that cultural diversity extends beyond language to include legal frameworks, news, and local content, noting that locally hosted content with local domain names remains crucial for cultural preservation.

Elisabeth Stewart Bradley from the International Trademark Association highlighted the particular challenges facing indigenous communities, whose traditional knowledge and cultural expressions are increasingly at risk of exploitation through AI systems without proper recognition or compensation. Peter Bruck from the World Summit Award provided a sobering perspective, arguing that social media and algorithmic content have actually worked against cultural preservation, with young people worldwide aspiring to become social media influencers rather than preserving their cultural heritage.

The panelists agreed that preserving linguistic and cultural diversity requires political will, adequate funding, and keeping this issue high on policy agendas, as access to information in one’s native language is a fundamental human right essential to respecting humanity’s full diversity.

Keypoints

**Major Discussion Points:**

– **Government and policy initiatives for language preservation**: Latvia’s proactive approach to digitalizing smaller languages, including subsidizing language resources, developing machine translation tools, and creating AI-powered public services in Latvian to ensure citizens can access government services in their native language.

– **Technical infrastructure and domain name systems**: The role of local internet registries and domain names in preserving cultural diversity, including locally hosted content, search engine optimization challenges, and the need to rebuild discoverability systems as AI-powered prompts potentially replace traditional search engines.

– **Indigenous rights and intellectual property protection**: The challenges of protecting traditional knowledge and cultural expressions from exploitation, particularly as AI compounds existing issues around cultural appropriation, and the need for balanced frameworks that respect indigenous communities while allowing respectful innovation.

– **The dark side of digital transformation**: A critical examination of how social media algorithms and big tech monopolization are actually eliminating cultural diversity rather than preserving it, with concerns about young people aspiring to become social media influencers rather than connecting with their cultural heritage.

– **AI as both threat and opportunity**: The dual nature of artificial intelligence in cultural preservation – while AI can help digitize endangered languages and provide translation services, there’s a risk that AI systems trained primarily on Western content will further marginalize minority cultures and languages.

**Overall Purpose:**

The discussion aimed to examine strategies for preserving cultural and linguistic diversity in the digital age, particularly with the rise of AI technologies, and to assess progress made since the original World Summit on the Information Society (WSIS) 20 years ago.

**Overall Tone:**

The discussion began with a measured, optimistic tone as speakers shared successful initiatives and technical solutions. However, the tone shifted notably toward concern and urgency, particularly with Professor Bruck’s “wake-up call” about the dark side of digital transformation. The conversation concluded on a sobering but determined note, emphasizing that cultural and linguistic diversity is a fundamental human right that requires continued vigilance and action to protect.

Speakers

– **Caroline Vuillemin**: Session moderator/chair

– **Gatis Ozols**: Deputy State Secretary for Digital Transformation of the Government of Latvia

– **Pierre Bonis**: Chief Executive Officer of AFNIC (French Internet Registry)

– **Elisabeth Stewart Bradley**: Representative from the International Trademark Association (INTA)

– **Peter A. Bruck**: Professor, World Summit Award founder/representative

– **Narendra Kumar Goyal**: Representative of an association with 74 partners and 48,500 members (specific organization not clearly mentioned)

– **Alfredo Ronchi**: Secretary General of EC Medici Framework, cooperates with UNESCO-IFAP initiative

**Additional speakers:**

None identified beyond the provided speakers names list.

# Leaders’ Talk: Preserving Cultural and Linguistic Diversity in the Digital Era

## Executive Summary

This Leaders’ Talk, moderated by Caroline Vuillemin as part of the WSIS+20 review focusing on Action Line C8 (cultural diversity, linguistic diversity and local content), brought together six distinguished speakers to examine the challenge of preserving cultural and linguistic diversity in an AI-dominated digital landscape. The discussion highlighted that while this challenge began with the rise of the internet, it has intensified with artificial intelligence systems predominantly trained on English-language content, representing over 90% of large language models.

Speakers presented diverse perspectives ranging from successful national digitization initiatives to critical assessments of technology’s impact on cultural preservation. The conversation emphasized that access to information in one’s native language constitutes a fundamental human right essential to respecting humanity’s full diversity.

## Opening Context and Framework

Caroline Vuillemin established that languages and culture encompass ways of seeing the world, representing ourselves, and understanding each other beyond mere words. She emphasized that preserving cultural diversity requires anticipation, political will, adequate funding, and keeping this issue high on policy agendas. The session was positioned within the broader WSIS+20 review context, examining progress made since the original summit and identifying emerging challenges in the digital transformation era.

## National Strategies and Initiatives

### Latvia’s Digital Language Development

Gatis Ozols, Deputy State Secretary for Digital Transformation of the Government of Latvia, presented Latvia’s comprehensive approach to digitalizing the Latvian language. Latvia began developing national policies 15-20 years ago to subsidize and develop language resources for digital tools, starting with machine translation initiatives. This early preparation has enabled Latvia to create AI-powered public services in Latvian, ensuring citizens can access government services in their native language.

Ozols emphasized a key principle: “Multilinguality and cultural diversity has to be viewed as core principle, not as a feature of gen AI.” He identified two paths being pursued – multinational companies training central global large language models, and the European Union approach supporting national and regional initiatives. He advocated for a dual approach, working with multinational companies while developing national resources.

### India’s Bhasani Programme

Narendra Kumar Goyal presented India’s Bhasani programme supporting 22 regional languages with 3 billion users, enabling speech translation across 21 languages. He outlined a comprehensive four-point framework:

1. **Use AI as a cultural ally**: Digitize languages through speech recognition and enable AI-powered translation tools

2. **Empower local voices**: Support content creators and promote intergenerational cultural exchange

3. **Tackle key challenges**: Address data gaps for lesser-known languages, ensure cultural sensitivity in AI design, and bridge the digital divide with adequate infrastructure and education

4. **Create an inclusive ecosystem**: Champion multi-language content creation and push for language equity in digital platforms

Goyal emphasized that “we cannot describe our culture and our land if we do not have a language,” highlighting that preserving culture means nurturing the spirit and identity of communities from within.

## Infrastructure and Technical Perspectives

### Internet Infrastructure and Cultural Frameworks

Pierre Bonis, Chief Executive Officer of AFNIC (French Internet Registry), expanded the discussion beyond language to broader cultural frameworks. He stressed that cultural diversity extends beyond language to include legal frameworks, news, and regulatory proximity that international companies may not respect. Local domain names and locally hosted content (.fr, .de, .sn) play a crucial role in cultural and linguistic diversity preservation.

Bonis raised a critical concern about AI’s impact on content discoverability: “If PROMPT replace search engines, in a way we are going to have to rebuild all the efforts that we did for 20 years in terms of search engine optimization, in terms of where are the sources, where can I check that this content is actually coming from a place.” He also mentioned ISOC France’s scrutiny of Facebook’s terms of use as an example of protecting local cultural and legal frameworks.

### Indigenous Rights and Intellectual Property

Elisabeth Stewart Bradley, representing the International Trademark Association (INTA), highlighted vulnerabilities facing indigenous communities in the digital age. She explained that traditional knowledge and cultural expressions face increased exploitation risks without proper recognition or compensation, with AI compounding existing issues around cultural appropriation.

Bradley outlined INTA’s five foundational principles for AI implementation in intellectual property protection, emphasizing the need for human oversight, transparency, and lawful access to data for rights enforcement. She argued for harmonized national frameworks to balance cultural preservation with innovation while respecting indigenous communities.

## Critical Assessment and Challenges

### Technology’s Impact on Cultural Preservation

Professor Peter A. Bruck from the World Summit Award provided a critical assessment of technology’s role in cultural preservation. He argued that current discussions focus too heavily on technical solutions while ignoring fundamental problems: “today we need to not just talk about the bright side, we need to talk about the dark side… if you are not willing to address that in terms of preserving culture and language in the digital era, you are just wrong.”

Bruck identified several key challenges:

– Young people worldwide aspire to become Instagram influencers rather than preserving cultural heritage

– Social media algorithms favor extremist emotional content and push misinformation

– The monopolization of user data by five big tech companies requires taxation to fund cultural preservation resources

– Without addressing hyperscale company monopolization, there will be insufficient resources for cultural preservation

### AI Systems and Cultural Representation

Alfredo Ronchi, Secretary General of EC Medici Framework, emphasized that different AI systems should be trained on different cultural documentation to avoid minoritizing creativities worldwide. He referenced the UNESCO-IFAP initiative and the atlas of living languages, stressing the importance of keeping humans at the center of technology to avoid losing cultural identity.

Ronchi argued that current AI systems predominantly reflect Western cultural perspectives, risking the marginalization of non-Western creativities and intelligences.

## Key Themes and Challenges

### AI as Both Threat and Opportunity

The discussion revealed AI’s dual nature in cultural preservation:

**Threats identified:**

– Large language models over 90% trained in one language create representation gaps

– AI systems predominantly reflect Western cultural perspectives

– Replacement of search engines with AI prompts threatens content discoverability

– Current AI training data lacks adequate representation of diverse cultures

**Opportunities recognized:**

– AI can digitize endangered languages and build living dictionaries

– Machine translation tools can provide native language access to services

– AI-powered systems can enable universal translation for international communication

– Technology can support intergenerational cultural exchange

### Unresolved Technical and Policy Challenges

Several critical challenges require further attention:

– Rebuilding search optimization efforts as AI prompts replace traditional search

– Addressing data gaps for lesser-known languages in AI training

– Developing harmonized frameworks balancing cultural preservation with innovation

– Creating funding mechanisms for cultural preservation initiatives

– Protecting indigenous rights and traditional knowledge from exploitation

## Conclusion

The Leaders’ Talk demonstrated that preserving cultural and linguistic diversity in the digital era requires coordinated efforts combining technical innovation, policy reform, and cultural awareness. While speakers presented different approaches – from collaborative engagement with multinational companies to critical confrontation of existing systems – they shared recognition that cultural and linguistic diversity represents a fundamental human right.

As Caroline Vuillemin concluded, preserving linguistic diversity and cultural heritage requires sustained political will, adequate funding, and continued vigilance to ensure technological advancement serves humanity’s full diversity. The session’s emphasis on diversity and inclusion as “basic human rights” and the need to respect “humanity in its globality” provides a foundation for future collaboration. A follow-up session was scheduled for Friday, with participants invited to join the World Summit Award global community for continued engagement on these critical challenges.

Caroline Vuillemin: Welcome everybody to this Leaders’ Talk on Local to Global, Preserving Culture and Language in a Digital Era. This session will be about inclusion and representation of all human diversity in today’s digital world. This is not a new question. It started more than 20 years ago with the rise of the Internet, but today in a truly digital world and with the rise of artificial intelligence, the question is more important than ever. So we will hear from various perspectives this afternoon on how to preserve that diversity if it is indeed important and a shared ambition and vision. I’d like to start with Mr. Ossorz, the Deputy State Secretary for Digital Transformation of the Government of Latvia. I have a question for you, sir. While Latvian is one of the smaller languages facing unique challenges in the digital age, the country of Latvia has emerged as a front-runner in advancing research and development for other underrepresented languages across the EU. Could you please describe your country’s strategy and the key initiatives aimed at digitalizing the Latvian language to ensure its comprehensive inclusion in AI-driven technologies?

Gatis Ozols: Hello. Thank you, madam, and it’s a pleasure to be here and share our international experience regarding this. When we speak now about AI and multilinguality and different cultures in the AI session, we already speak about that currently the large language models that are largely used by the community, by consumer products, largely are more than 90% trained by one language. Actually, we recognize this important aspect as being a small and compact country, so we have to take care of our language and digital environment early on. It was already, I suppose, some 15, 20 years ago when it started with machine translation and such tools, so we put in a high position in our national policy, and we started to subsidize and we started to pay special attention to develop those national language resources to train digital tools for the Latvian language. Now we see it as an important aspect not only for the Latvian language, but actually for most of the languages in the world which are underrepresented on digital environment, also on AI tools. So how we approach it, initially we developed and facilitated programs nationally to develop language resources on Latvian language. It was like initial to build the base, so to train initially the machine translation to ensure that our Latvian citizens can use internet, which is mostly English-driven, and our own natural language for e-commerce, it was also very essential. Then of course it was also essential to access public services, so initially for European countries, it’s like more than 27 countries, you have to access in different languages, but then when we developed this language technologies, we saw it also helps for accessibility of public services not only, but also in internet, like text-to-speech, speech-to-text, it’s ensure accessibility. Also for public services, the national platform for AI assistance for the citizens, so they can approach the government digitally, they’ve been answered in the Latvian language. So this has allowed us to ensure this multilinguality and accessibility of public services and public resources, and also to ensure that those resources are available for the citizens. Further now with AI, I think it becomes even more critical to ensure this multi-language and multicultural representation, as AI becomes the first of the main way how we interact with different services and consumer products, and I would like to focus briefly on three aspects to ensure this. First one, multilinguality as a core principle, access to language resources and access to compute and expertise. Multilinguality and cultural diversity has to be viewed as core principle, not as a feature of gen AI. It’s our position and our view nationally. We see that there are different paths that have been taken, industry part, it’s taken by companies, multinationals that train central global LLMs, and there’s another part, or the way how to approach this, is European Union approach, that we have this national regional initiatives to incentivise and support development of large language models, in our case for EU countries, and provide access to compute. So we have to work on both of those dimensions together with those multinational companies and also to develop our national resources to ensure that all the languages and cultures are represented nationally.

Caroline Vuillemin: Thank you. So we see that it is about setting priorities, anticipating and having political will to preserve this diversity and languages. I’d like to see now another perspective from Pierre Bonis, the Chief Executive Officer of AFNIC. How do you perceive the importance of preserving cultural and linguistic diversity through the lens of the French Internet Registry?

Pierre Bonis: Thank you very much, first of all. Thank you for the invitation. I’m sorry to have to speak English, and I’m hoping that my English or British counterparts will not end this week with a very big loss of cultural diversity, which is speaking globish. So we started, of course, 20 years ago, because this is WSIS plus 20, with this Action Line C8, cultural diversity, linguistic diversity and local content. I think it was very wise to mix these different aspects, and especially the local content. And if you look at the challenges that we have, and we still have, in the discoverability of the different contents in different languages and different cultures, the role of the locally host and made content is still very relevant. In that regard, from the French Registry and from other registries, .fr, .de, .sn, etc., the fact that the contents may be hosted locally and may be addressed with local domain names makes sense also. More broadly, we really think that the problem is not only the language, of course it’s very important, access to your own language online, and sometimes the domain name can help, sometimes the search engine can help, but this is also culture, this is news, this is legal framework. I’ll give you an example. In France, ISOC France scrutinized the Facebook terms of use in France. And just so that this is another part of cultural approach, I mean, the legal framework that can be imposed by an international company may not be exactly the one that we are used to in our various countries. So this cultural diversity goes far beyond the pure language approach. It’s really a question of legal, regulatory, and proximity framework. I would end with just quoting, and Mr. Minister, you said that very well, that we have new challenges and opportunity with AI. But one of the challenges we face is that if PROMPT replace search engines, in a way we are going to have to rebuild all the efforts that we did for 20 years in terms of search engine optimization, in terms of where are the sources, where can I check that this content is actually coming from a place. So I think, I mean, this is not, this is just being aware of it. But being aware of it is making sure that the AI is not going to… Once again, having all the work to do to make sure that our cultures, our diversity is seen on the Internet. Thank you.

Caroline Vuillemin: Thank you very much. As you said, languages and culture are not only words. There are also ways of seeing the world, ways of representing ourselves, ways of understanding each other. So it’s very important to take this into account with the technology. I’d like now to turn to Elisabeth Stewart Bradley from the International Trademark Association. So another point of view still. Your association is dedicated to the protection of intellectual property to foster consumer trust, economic growth and innovation. And it is also committed to building a better society through brands. How does the protection of linguistic identity and local content fit into your organization’s mission?

Elisabeth Stewart Bradley: Thank you for this question. I’m sorry for the glasses, but I really am having trouble here. So first, I would like to thank ITU Secretary General Doreen Bogdan-Martin, the WSIS sponsors and organizers, and my fellow high-level panelists. It’s an honor to be here with you today. INTA’s mission seeks sustainable answers to global challenges. This is complementary to the SDGs as our work focuses primarily on SDG 9, industry, innovation and infrastructure. Last year, our intervention emphasized the importance of enhancing the availability of domain names in local languages. While domain names remain top of mind, there is now an even greater focus on how AI affects the preservation of culture and language. For example, the protection of indigenous rights has long presented challenges to both brand owners and indigenous communities. And AI is now compounding that. Traditional knowledge and cultural expressions have been passed down through generations. But in many cases, they have been exploited and appropriated without proper recognition or compensation. This leads to the loss of control of indigenous people over their cultural heritage. For brand owners, this presents challenges in terms of respecting the rights of indigenous communities while also wanting to adopt elements of their culture into products, services, and marketing strategies in reasonable and respectful ways. INTA has a dedicated indigenous rights committee that includes a globally diverse membership. And we actively participate in discussions at WIPO. Our goal within the global community is to harmonize and balance how indigenous culture and language may be protected and incorporated in the development of new products and services. Notably, there have been positive advancements in recognition and safeguarding of indigenous rights, including legal reforms and greater awareness of indigenous issues. However, challenges persist, like the lack of harmonized national frameworks to ensure the proper balance between cultural preservation and innovation. Rapid advancements in technology have raised concerns about indigenous heritage use by AI, lack of representation, and lack of consultation. In terms of the implementation of AI to help protect IP, including indigenous rights, INTA supports policies based on five foundational principles that include recognizing human versus machine contributions, final decisions on the granting or revocation of rights should be subject to human oversight, rights holders should be able to obtain lawful access to data for the purpose of enforcing IP rights, customers should know the source of information received by AI, and transparency should be balanced with the need to protect proprietary information. INTA fosters the exchange of global perspectives by its thousands of members from around the world. INTA recently adopted AI-based technology to provide universal translation services for participants at our meetings. Our most recent annual meeting hosted more than 10,000 participants. Until the implementation of AI-based translation programs, organizations such as INTA could not afford translation services. Providing greater access through translation allows participants to return to their countries with powerful information to help preserve their valuable inventions. In closing, I’d just like to say that WSIS enables the spirit of multi-stakeholder engagement. Thank you so much.

Caroline Vuillemin: Thank you. And thank you for highlighting the challenge of the national frameworks versus the global approach. At the current era of crisis in multilateralism, I think national frameworks still have to exist, but to be comprehensive and coordinated among various nations. We will now listen to Professor Brook from the World Summit Award, and maybe look back 20 years ago when you initiated at the first WSIS conference in Geneva the World Summit Award. How has the challenge of preserving culture and language in the digital era changed since then?

Peter A. Bruck: Thank you very much, Caroline. That’s a wonderful question, which leads me to say that in 2003 we looked at the information and communication technologies with optimism. And today we need to not just talk about the bright side, we need to talk about the dark side. And only those who are addressing the dark side and the issues are actually credible interlocutors for where we should go in the future, because we cannot parrot again what we have talked about and the way we have talked about in 2003 about the WSIS Action Plan. I had lunch today with three people, one person from Jordan, another one from Malaysia, and the other one from Cyprus. And they all know that their children and their grandchildren, female, 14 years old, have as an idea for where they want to go in their lives to be an Instagram influencer. And I say to you, everybody in this room, if you are not willing to address that in terms of preserving culture and language in the digital era, you are just wrong in terms of this place. You have to address it, we have to address it. The huge achievement of social media has in fact turned against preserving cultural heritage. And cultural diversity is not just reduced, but it is eliminated around the world in many different ways. And I think that if we are not looking at what algorithmic media do in terms of favoring not just this kind of fantasies of young women, but also in terms of democracy by favoring extremist emotional, cultural, and political content, and push nonsense to people, lies, hate speech, and fake news, we are just missing the opportunity of looking at business plus 20. User data are collected from every corner of the world, and they are monopolized by five big tech companies. And if we are not addressing the monopolization of that, which is hyperscaling, by taxing those hyperscale companies, we are just not having the resources for cultural preservation and diversity. The World Summit Award was created actually in 2003, ironically, to identify the richness and the plurality, the diversity of high quality contents around the world. And we were looking at that time at ICT as being a means to do it. Today, I have to acknowledge the needs for the World Summit Award, and its global community of creative people who look for a positive social impact is not lesser, but greater. Because we are losing diversity, we are losing our cultural heritage, and we are not preserving it. So, from my point of view, I invite each and everyone in this room to join us in terms of this global community, and also to act in terms of making WSA also a follow-on activity to WSIS 2030. Thank you. I hope I am in time.

Caroline Vuillemin: Yes, perfect.

Peter A. Bruck: I can see that there are some people who are on Instagram and who clap still. Thank you very much.

Caroline Vuillemin: Thank you, and thank you for that wake-up call, which should not be a wake-up call, because I think we all have the same reality in front of us in terms of losing diversity in the last 20 years, which is not what the spirit of WSA is.

Narendra Kumar Goyal: Thank you very much. And I’m happy that before me our panelists said two important things. Social media influencing and changing our life. The new generation is forgetting the old culture, our language, our heritage. And five tech companies trying to do everything for all of us. Led by these things, my association has 74 partners with a large base of 48,500 members. So we discussed this ourselves, what is the way out to preserve the cultural heritage. We came out with two statements and four points in brief. The statement says we cannot describe our culture and our land if we do not have a language. And preserving culture means nurturing the spirit and identity of communities from within. Now there are four things. One is use AI as a cultural ally. Digitize intense languages via speech recognition and transcription. Enable access with AI powered translation and language learning tools. Build curriculum and living dictionaries to preserve oral traditions. Second point was empower local voices. Support the creators of native languages with the digitalized tools, storytelling, and next promote intergenerational exchange through technical enabled mentorship. Celebrate culture online, customs, folk fare, musics. Third point is tackle the challenges. Address data gaps for developing lesser known languages. Ensure cultural sensibility via AI design. Bridge the digital divide with infrastructure and education. And the last topic is create inclusive ecosystem. In this champion multi-language content across media platforms. Push for language equity in the technical tools, transforms, fund preservation projects at the government and grassroot level. Within the next 30 minutes I want to share the Indian experience. India has 22 regional languages. Our government made a program called Bhasani, Indian languages. And we have, as a man there is a speech by our leader. It is transferred into 21 languages across the country. There are 3 billion users as of now. Thank you very much.

Caroline Vuillemin: Thank you very much. Impressive figures from a Swiss citizen with only four languages in our country. Thank you for your contribution. Last but not least I want to turn now to Prof. Alfredo Ronchi. You’re the Secretary General of EC Medici Firmwork. So maybe a last question to give us hope and perspective. Can the digital age jeopardize cultural and linguistic diversity?

Alfredo Ronchi: Okay, thank you Madam Chair. Yes, so good afternoon first of all. So yes, this is a question that was posed a long time ago. As you introduced the session before, it’s something that basically coupled with the explosion of the use of the internet that boosted the idea of globalization with the risk to jeopardize cultural models, cultural identities, and of course languages. I used to cooperate with UNESCO-IFAP initiative and this was one of the key topics to be discussed in the meeting. So the fact that the atlas of living languages was diminishing the size every year due to the disappearance of a number of languages, minoritized so-called languages, because they were not able to be both spoken and then represented on the internet. And of course this idea to get toward uniformity to provide to young generation a standard model that is almost equal all over the world may probably pose the risk to lose our identities in terms of languages that is tightly connected with culture. I am very, let’s say, fond of studying different languages because it’s the best way to understand better some culture, some population, because there’s this tight relation even in the way in which it’s structured, the language. And so it’s for sure the risk to lose some relevant values. And this is really a pity in the set of the culture and languages. But nowadays there’s another technology that comes back after 30 years more or less. In the 80s it was considered one of the monster artificial intelligence, but there was no chance to keep in touch with it at that time because it was too abstract. Let’s say expert system and fuzzy logic. Nowadays thanks to some products like GPT and similar things, people can experience this so they lower in order to enter this technology is very low. Anyone can try it. And the risk again, because we fit such kind of system mainly with some specific set of documents, let’s say some Western culture basically, there’s the risk to lose some other intelligences. So last year we spoke and within some of the sessions at the WISIS, mainly managed by UNESCO, to the idea to create different AI systems fed with different culture, different documentation, in order to do not risk to have minoritized creativities around the world. And this of course is tightly connected with the road to extend, let’s say, the ability to create content in such kind of specific culture and so on. So it is very complicated things. But the point is to try to keep humans in the center and do not become slave of technology, losing our identity. Thank you very much.

Caroline Vuillemin: Thank you. Well, thank you all for your contributions and for keeping with the time. I know it’s very difficult to go into deep thoughts in only three minutes in this topic that is so important and where we would all have hoped to speak in our own language and be very diversified. But maybe next time. As a conclusion, just to wrap up, diversity and inclusion of course are not only values. It is a basic human right. Each individual on earth should have access to content, to information regarding its government, regarding news, regarding economic activities in his or her own maternal language. And it’s not only for commercial incentives and human rights, but it’s really to respect humanity in its globality. So we heard it’s about anticipation, political will, funding, and keeping this high on the agenda. I hope the conclusions of this summit will drive policymakers in that direction. A final summary of the session will be provided and it will take place on Friday, this Friday the 11th at 3 p.m. So please join us to listen to the final conclusion then. Thank you.

Agreement points

AI systems predominantly trained in one language create representation gaps for smaller languages

Speakers

– Gatis Ozols
– Alfredo Ronchi

Arguments

Large language models are over 90% trained in one language, creating representation gaps for smaller languages

Different AI systems should be fed with different cultural documentation to avoid minoritizing creativities worldwide

Summary

Both speakers recognize that current AI systems are biased toward dominant languages/cultures, with Ozols providing specific statistics about 90% single-language training and Ronchi emphasizing the Western cultural bias in AI training data

Topics

Cultural diversity | Digital standards | Multilingualism

Government-led national strategies are essential for language preservation

Speakers

– Gatis Ozols
– Narendra Kumar Goyal

Arguments

Latvia developed early national policies to subsidize and develop language resources for digital tools, starting with machine translation 15-20 years ago

India implemented the Bhasani program supporting 22 regional languages with 3 billion users, enabling speech translation across 21 languages

Summary

Both speakers present successful examples of government-initiated programs for language digitization, with Latvia’s early machine translation efforts and India’s massive Bhasani program demonstrating the effectiveness of national strategies

Topics

Cultural diversity | Multilingualism | Development

Cultural preservation extends beyond language to encompass broader frameworks

Speakers

– Pierre Bonis
– Caroline Vuillemin

Arguments

Cultural diversity extends beyond language to include legal frameworks, news, and regulatory proximity that international companies may not respect

Languages and culture encompass ways of seeing the world, representing ourselves, and understanding each other beyond just words

Summary

Both speakers emphasize that cultural preservation involves comprehensive worldviews, legal systems, and ways of understanding reality, not just linguistic translation

Topics

Cultural diversity | Legal and regulatory | Digital identities

Technology threatens cultural diversity and requires human-centered approaches

Speakers

– Peter A. Bruck
– Alfredo Ronchi

Arguments

Social media algorithms favor extremist emotional content and push misinformation, working against cultural heritage preservation

Technology should keep humans at the center to avoid becoming slaves to technology and losing cultural identity

Summary

Both speakers warn about technology’s potential to undermine cultural diversity and emphasize the need to maintain human agency and control over technological systems

Topics

Cultural diversity | Content policy | Digital identities

Similar viewpoints

Both speakers advocate for proactive, positive approaches to AI implementation in cultural preservation, viewing AI as a tool that can support rather than threaten linguistic diversity when properly implemented with strong national strategies

Speakers

– Gatis Ozols
– Narendra Kumar Goyal

Arguments

Multilinguality and cultural diversity must be viewed as core principles, not features, requiring both multinational company cooperation and national resource development

AI should be used as a cultural ally through digitizing languages, enabling translation tools, and building living dictionaries

Topics

Cultural diversity | Multilingualism | Digital standards

Both speakers emphasize the need for systematic approaches to ensure diverse cultural perspectives are properly represented and protected in technological and legal frameworks, rather than being marginalized by dominant systems

Speakers

– Elisabeth Stewart Bradley
– Alfredo Ronchi

Arguments

Brand owners need harmonized national frameworks to balance cultural preservation with innovation while respecting indigenous communities

Different AI systems should be fed with different cultural documentation to avoid minoritizing creativities worldwide

Topics

Cultural diversity | Legal and regulatory | Intellectual property rights

Both speakers express concern about how algorithmic systems and AI are undermining decades of work in making diverse content discoverable and accessible, potentially erasing cultural diversity achievements

Speakers

– Pierre Bonis
– Peter A. Bruck

Arguments

AI replacing search engines with prompts threatens 20 years of work in search engine optimization and content discoverability

Social media algorithms favor extremist emotional content and push misinformation, working against cultural heritage preservation

Topics

Cultural diversity | Content policy | Digital standards

Unexpected consensus

Need for taxation of big tech companies to fund cultural preservation

Speakers

– Peter A. Bruck

Arguments

User data monopolization by five big tech companies requires taxation to fund cultural preservation resources

Explanation

While only Bruck explicitly mentioned this, his argument about taxing hyperscale companies to fund cultural preservation represents a concrete policy solution that aligns with other speakers’ concerns about tech company dominance, though no other speaker directly addressed this economic approach

Topics

Cultural diversity | Taxation | Digital business models

AI can serve as both threat and ally to cultural preservation

Speakers

– Gatis Ozols
– Narendra Kumar Goyal
– Alfredo Ronchi

Arguments

Multilinguality and cultural diversity must be viewed as core principles, not features, requiring both multinational company cooperation and national resource development

AI should be used as a cultural ally through digitizing languages, enabling translation tools, and building living dictionaries

Different AI systems should be fed with different cultural documentation to avoid minoritizing creativities worldwide

Explanation

Despite concerns about AI’s threats to cultural diversity, multiple speakers converged on the idea that AI can be harnessed positively for cultural preservation when properly designed and implemented with diverse training data and multilingual principles

Topics

Cultural diversity | Digital standards | Multilingualism

Overall assessment

Summary

Speakers demonstrated strong consensus on the fundamental importance of cultural and linguistic diversity preservation, the inadequacy of current AI systems in representing diverse cultures, the need for government-led national strategies, and the recognition that cultural preservation extends beyond language to encompass worldviews and legal frameworks

Consensus level

High level of consensus on problem identification and general principles, with speakers agreeing that current technological systems threaten cultural diversity and that proactive, multi-stakeholder approaches are needed. The consensus suggests a mature understanding of the challenges and points toward coordinated policy responses combining national strategies, international cooperation, and human-centered technology design.

Different viewpoints

Approach to AI development for cultural preservation

Speakers

– Gatis Ozols
– Alfredo Ronchi

Arguments

Multilinguality and cultural diversity must be viewed as core principles, not features, requiring both multinational company cooperation and national resource development

Different AI systems should be fed with different cultural documentation to avoid minoritizing creativities worldwide

Summary

Ozols advocates for a dual approach working with multinational companies while developing national resources, viewing multilinguality as a core principle. Ronchi argues for creating separate AI systems trained on different cultural documentation to avoid Western cultural dominance, suggesting a more fragmented approach to AI development.

Topics

Cultural diversity | Digital standards | Multilingualism

Assessment of technology’s impact on cultural diversity

Speakers

– Peter A. Bruck
– Narendra Kumar Goyal

Arguments

Social media algorithms favor extremist emotional content and push misinformation, working against cultural heritage preservation

AI should be used as a cultural ally through digitizing languages, enabling translation tools, and building living dictionaries

Summary

Bruck presents a pessimistic view of technology, arguing that social media algorithms actively work against cultural preservation by promoting extremist content. Goyal takes an optimistic stance, viewing AI as a potential ally that can be harnessed for cultural preservation through practical applications.

Topics

Cultural diversity | Content policy | Multilingualism

Role of big tech companies in cultural preservation

Speakers

– Peter A. Bruck
– Gatis Ozols

Arguments

User data monopolization by five big tech companies requires taxation to fund cultural preservation resources

Multilinguality and cultural diversity must be viewed as core principles, not features, requiring both multinational company cooperation and national resource development

Summary

Bruck advocates for a confrontational approach toward big tech companies, calling for taxation to fund cultural preservation. Ozols suggests a collaborative approach, emphasizing the need to work with multinational companies alongside national resource development.

Topics

Cultural diversity | Digital business models | Taxation

Unexpected differences

Optimism vs. pessimism about digital technology’s role

Speakers

– Peter A. Bruck
– Narendra Kumar Goyal

Arguments

Young people worldwide aspire to become Instagram influencers, representing a threat to cultural diversity that must be addressed

India implemented the Bhasani program supporting 22 regional languages with 3 billion users, enabling speech translation across 21 languages

Explanation

This disagreement is unexpected because both speakers are addressing the same fundamental challenge of cultural preservation, yet they have dramatically different assessments of technology’s potential. Bruck’s pessimistic view contrasts sharply with Goyal’s success story, suggesting different regional experiences or philosophical approaches to technology adoption.

Topics

Cultural diversity | Digital identities | Multilingualism

Overall assessment

Summary

The main areas of disagreement center on approaches to AI development, assessment of technology’s impact on culture, and strategies for dealing with big tech companies. While all speakers agree on the importance of cultural preservation, they differ significantly on whether to work with or against existing tech infrastructure.

Disagreement level

Moderate disagreement with significant implications. The disagreements reflect fundamental philosophical differences about technology’s role in society and appropriate policy responses. These differences could lead to fragmented approaches to cultural preservation, potentially undermining coordinated global efforts. However, the shared commitment to cultural diversity provides a foundation for potential compromise and collaboration.

Partial agreements

Similar viewpoints

Both speakers advocate for proactive, positive approaches to AI implementation in cultural preservation, viewing AI as a tool that can support rather than threaten linguistic diversity when properly implemented with strong national strategies

Speakers

– Gatis Ozols
– Narendra Kumar Goyal

Arguments

Multilinguality and cultural diversity must be viewed as core principles, not features, requiring both multinational company cooperation and national resource development

AI should be used as a cultural ally through digitizing languages, enabling translation tools, and building living dictionaries

Topics

Cultural diversity | Multilingualism | Digital standards

Both speakers emphasize the need for systematic approaches to ensure diverse cultural perspectives are properly represented and protected in technological and legal frameworks, rather than being marginalized by dominant systems

Speakers

– Elisabeth Stewart Bradley
– Alfredo Ronchi

Arguments

Brand owners need harmonized national frameworks to balance cultural preservation with innovation while respecting indigenous communities

Different AI systems should be fed with different cultural documentation to avoid minoritizing creativities worldwide

Topics

Cultural diversity | Legal and regulatory | Intellectual property rights

Both speakers express concern about how algorithmic systems and AI are undermining decades of work in making diverse content discoverable and accessible, potentially erasing cultural diversity achievements

Speakers

– Pierre Bonis
– Peter A. Bruck

Arguments

AI replacing search engines with prompts threatens 20 years of work in search engine optimization and content discoverability

Social media algorithms favor extremist emotional content and push misinformation, working against cultural heritage preservation

Topics

Cultural diversity | Content policy | Digital standards

Key takeaways

Cultural and linguistic diversity preservation is a fundamental human right, not just a commercial or technological feature

Multilinguality must be treated as a core principle in AI development, requiring both cooperation with multinational companies and development of national resources

Early anticipation and political will are crucial – countries like Latvia started developing language resources 15-20 years ago with machine translation

AI poses both opportunities and threats: it can serve as a cultural ally through digitization tools, but current large language models are over 90% trained in one language

Social media and algorithmic content are actively working against cultural diversity by promoting homogenization and extremist content

Local internet infrastructure (domain names, hosting) plays a vital role in preserving cultural diversity beyond just language translation

The digital divide and data monopolization by five big tech companies requires systemic solutions including taxation to fund cultural preservation

Indigenous rights and traditional knowledge face increased exploitation risks with AI development, requiring harmonized national frameworks

Resolutions and action items

A final summary session scheduled for Friday the 11th at 3 p.m. to provide conclusions

Invitation to join the World Summit Award global community as a follow-on activity to WSIS 2030

Need for policymakers to keep cultural and linguistic diversity high on the agenda based on summit conclusions

Development of different AI systems fed with different cultural documentation to avoid minoritizing creativities

Implementation of AI-based technology for universal translation services in international meetings and organizations

Unresolved issues

How to effectively address the influence of social media on young people’s cultural aspirations (Instagram influencer phenomenon)

Lack of harmonized national frameworks to balance cultural preservation with innovation

How to rebuild content discoverability efforts if AI prompts replace search engines

Addressing data gaps for lesser-known languages in AI training

How to ensure cultural sensitivity in AI design and implementation

Bridging the digital divide with adequate infrastructure and education

How to effectively tax hyperscale tech companies to fund cultural preservation resources

Suggested compromises

Working on both dimensions: cooperating with multinational companies while developing national language resources

Balancing transparency in AI systems with the need to protect proprietary information

Finding ways for brand owners to respectfully incorporate indigenous cultural elements while protecting indigenous rights

Using technology as a tool while keeping humans at the center to avoid losing cultural identity

Developing both global AI solutions and region-specific cultural AI systems simultaneously

Multilinguality and cultural diversity has to be viewed as core principle, not as a feature of gen AI… there are different paths that have been taken, industry part, it’s taken by companies, multinationals that train central global LLMs, and there’s another part… European Union approach, that we have this national regional initiatives

Speaker

Gatis Ozols

Reason

This comment reframes the entire discussion by distinguishing between treating linguistic diversity as an optional feature versus a fundamental principle. It also introduces the critical tension between centralized global AI models versus decentralized national/regional approaches, highlighting a key strategic choice facing policymakers.

Impact

This established the foundational framework for the entire discussion, setting up the core tension between global tech companies’ approaches and national preservation efforts. Subsequent speakers built upon this dichotomy, with Pierre Bonis discussing local content hosting and Elisabeth Stewart Bradley addressing the need for harmonized national frameworks.

If PROMPT replace search engines, in a way we are going to have to rebuild all the efforts that we did for 20 years in terms of search engine optimization, in terms of where are the sources, where can I check that this content is actually coming from a place

Speaker

Pierre Bonis

Reason

This insight reveals a profound but often overlooked consequence of AI advancement – that the shift from search engines to AI prompts could undo decades of work in making diverse content discoverable. It highlights how technological progress can inadvertently create new barriers to cultural preservation.

Impact

This comment shifted the discussion from celebrating AI’s potential to acknowledging its disruptive risks. It introduced urgency to the conversation and influenced later speakers to address the ‘dark side’ of technology, particularly Peter Bruck’s more critical assessment.

Today we need to not just talk about the bright side, we need to talk about the dark side… if you are not willing to address that in terms of preserving culture and language in the digital era, you are just wrong… The huge achievement of social media has in fact turned against preserving cultural heritage

Speaker

Peter A. Bruck

Reason

This was a powerful wake-up call that challenged the panel’s relatively optimistic tone. Bruck directly confronted the reality that social media and algorithmic content have actively undermined cultural diversity, using the concrete example of young girls aspiring to be Instagram influencers as evidence of cultural homogenization.

Impact

This comment created a dramatic tonal shift in the discussion, moving from technical solutions and policy frameworks to a more critical examination of technology’s actual impact. It forced subsequent speakers to acknowledge the severity of the challenges and influenced Caroline Vuillemin’s acknowledgment that this ‘should not be a wake-up call’ because the reality is already evident.

User data are collected from every corner of the world, and they are monopolized by five big tech companies. And if we are not addressing the monopolization of that, which is hyperscaling, by taxing those hyperscale companies, we are just not having the resources for cultural preservation and diversity

Speaker

Peter A. Bruck

Reason

This comment connects cultural preservation directly to economic power structures, arguing that without addressing tech monopolization and creating funding mechanisms through taxation, cultural preservation efforts will remain under-resourced. It’s a systemic analysis that goes beyond technical solutions.

Impact

This introduced an economic dimension to the discussion that hadn’t been explicitly addressed before, suggesting that cultural preservation requires not just technical solutions but fundamental changes to how tech companies are regulated and taxed. It influenced Narendra Kumar Goyal’s subsequent emphasis on government funding and grassroots support.

We cannot describe our culture and our land if we do not have a language. And preserving culture means nurturing the spirit and identity of communities from within

Speaker

Narendra Kumar Goyal

Reason

This philosophical insight captures the fundamental relationship between language and cultural identity, emphasizing that preservation must come from within communities rather than being imposed externally. It provides a humanistic counterpoint to the technical discussions.

Impact

This comment grounded the discussion in fundamental human values and shifted focus toward community-driven solutions. It influenced the moderator’s final emphasis on diversity and inclusion as ‘basic human rights’ and helped frame the conclusion around respecting ‘humanity in its globality.’

Overall assessment

These key comments transformed what could have been a purely technical discussion about AI and language tools into a nuanced examination of power structures, cultural values, and systemic challenges. The progression from Ozols’ strategic framework, through Bonis’ technical concerns, to Bruck’s critical wake-up call, and finally to Goyal’s humanistic grounding created a comprehensive dialogue that acknowledged both opportunities and threats. The comments collectively shifted the discussion from optimistic problem-solving to realistic assessment of the scale of challenges, ultimately leading to Caroline Vuillemin’s conclusion that framed linguistic diversity as a fundamental human right requiring political will and sustained funding.

How to rebuild search engine optimization efforts and source verification systems when AI prompts replace traditional search engines

Speaker

Pierre Bonis

Explanation

This addresses a critical challenge where 20 years of work on search engine optimization and content discoverability may need to be rebuilt as AI systems change how people access information

How to develop harmonized national frameworks that balance cultural preservation with innovation, particularly for indigenous rights protection

Speaker

Elisabeth Stewart Bradley

Explanation

The lack of consistent international frameworks creates challenges in protecting indigenous cultural heritage while allowing respectful innovation and commercial use

How to address the monopolization of user data by five big tech companies and implement taxation of hyperscale companies to fund cultural preservation

Speaker

Peter A. Bruck

Explanation

This tackles the fundamental issue of resource concentration and funding mechanisms needed to support cultural diversity initiatives globally

How to address data gaps for developing AI systems that support lesser-known and minoritized languages

Speaker

Narendra Kumar Goyal

Explanation

Many languages lack sufficient digital data to train AI systems, creating barriers to their preservation and digital inclusion

How to create different AI systems fed with diverse cultural documentation to avoid minoritizing non-Western creativities and intelligences

Speaker

Alfredo Ronchi

Explanation

Current AI systems are predominantly trained on Western content, risking the loss of diverse cultural perspectives and ways of thinking

How to counter the homogenizing effects of social media algorithms that favor extremist content and eliminate cultural diversity

Speaker

Peter A. Bruck

Explanation

Social media platforms are actively reducing cultural diversity through algorithmic promotion of uniform content and extremist material

How to ensure cultural sensitivity in AI design and development processes

Speaker

Narendra Kumar Goyal

Explanation

AI systems need to be designed with cultural awareness to avoid perpetuating biases or misrepresenting cultural content

How to scale successful national language digitization programs like Latvia’s and India’s Bhasani to other countries and languages

Speaker

Gatis Ozols and Narendra Kumar Goyal

Explanation

Understanding how to replicate successful models could help preserve more languages globally in the digital age