Media and Education for All: Bridging Female Academic Leaders and Society towards Impactful Results

7 Jul 2025 17:00h - 18:00h

Media and Education for All: Bridging Female Academic Leaders and Society towards Impactful Results

Session at a glance

Summary

This discussion focused on media and education accessibility, featuring presentations from female academic leaders who are bridging university research with real-world applications through AI-powered solutions. The session was moderated by Dr. Anita Lamprecht from the Diplo Foundation and organized by the AXS-CAT network from Universitat Autonoma de Barcelona, which aims to transfer academic knowledge to industry and society.


Professor Ana Matamala presented the ENACT project, which develops easy-to-understand news content across web, radio, and television platforms to serve diverse audiences including people with intellectual disabilities, migrants, and language learners. The project involves European broadcasters and has identified 13 organizations already producing simplified news content, though challenges remain in balancing accessibility with journalistic principles. Professor Pilar Orero discussed three AI-enhanced projects: Mosaic, which creates a searchable hub for European public broadcaster archives; Alfie, which addresses biases in media and AI models; and a climate change communication project that makes scientific data more accessible to citizens.


Dr. Estella Oncins presented the Inclusivity project, an Erasmus Plus initiative exploring virtual reality in education, particularly for neurodiverse students. Initial results showed positive reception from both teachers and students, with emphasis on hands-on activities and the need for comprehensive training tools. Dr. Mireia Farrus introduced Scribal, a real-time transcription and translation system designed for university environments, specifically supporting Catalan language and dialects while ensuring data privacy. Dr. Mar Gutierrez Colon shared two accessibility projects: a gamified reading app for Kiswahili-speaking children in Kenya and an English proficiency test designed for students with special educational needs.


The discussion concluded with audience questions about challenges in teaching emerging technologies like AI in universities, particularly the difficulty of finding qualified instructors and the rapid pace of technological change. The speakers emphasized the importance of industry-academia collaboration and described digital accessibility as a form of educational justice that benefits all learners.


Keypoints

## Major Discussion Points:


– **Media Accessibility Through AI Technology**: Multiple speakers presented projects using AI to make media content more accessible, including easy-to-understand news (ENACT project), automated transcription and translation systems (Scribal), and enhanced subtitle creation for minority languages.


– **Educational Inclusion and Digital Accessibility**: Several projects focused on making education more inclusive, particularly the Inclusivity VR project for neurodiverse students, English proficiency testing for students with special educational needs, and gamified reading apps for underserved communities in Kenya.


– **Challenges in AI Implementation and Bias**: Discussion of significant barriers including lack of trained educators for new technologies, persistent bias in AI translation systems (particularly gender bias), and the gap between rapidly evolving industry knowledge and academic curriculum development.


– **Knowledge Transfer from Academia to Society**: The session emphasized bridging the gap between university research and real-world applications, with the AXS-CAT network facilitating knowledge transfer to industry and society to solve practical communication and accessibility problems.


– **Language Preservation and Minority Language Support**: Focus on supporting minority languages like Catalan through specialized transcription systems, and addressing the needs of multilingual communities including migrants and students with limited language proficiency.


## Overall Purpose:


The discussion aimed to showcase how female academic leaders are translating university research into practical solutions that address real-world accessibility and communication challenges, particularly through AI-powered tools that make media and education more inclusive for diverse populations.


## Overall Tone:


The discussion maintained a consistently professional and collaborative tone throughout. It was informative and solution-oriented, with speakers presenting their research projects efficiently within time constraints. The tone became more interactive and engaged during the Q&A session, where practical challenges were discussed openly, particularly around the difficulties of teaching emerging technologies and addressing AI bias. The moderator maintained an encouraging and appreciative tone, emphasizing the significance of having multiple female experts in one session.


Speakers

– **Anita Lamprecht**: Dr., Moderator, Diplo Foundation, Emerging tech expert and researcher


– **Ana Matamala**: Professor, Thomas University of Barcelona (mentioned as Universitat Autonoma de Barcelona), Leading the ENACT project on easy-to-understand news


– **Pilar Orero**: Professor, Expert in media accessibility and AI, Working on projects including Mosaic, Alfie, and climate change accessibility


– **Estel la Oncins**: Dr., Universitat Autònoma de Barcelona, Leading the Inclusivity project on virtual reality in education


– **Mireia Farrus**: Dr., University of Barcelona, Presenting the Scribal project for audio transcriptions in university teaching


– **Mar Gutierrez Colon**: Dr., Universitat Rovira Virgili, Working on accessibility projects including language learning apps and English proficiency tests for SEND students


– **Audience**: Various audience members asking questions


Additional speakers:


– **NK Goel**: Professor from India, Connected to 100,000 colleges and universities in India


– **Tamiris**: Works with fact-checking


Full session report

# Comprehensive Report: Media and Education Accessibility Through AI-Enhanced Solutions


## Introduction and Context


This discussion, moderated by Dr. Anita Lamprecht from the Diplo Foundation, brought together leading female academics from Spanish universities to showcase how university research is being translated into practical solutions for media and education accessibility. The session was organised by the AXS-CAT network from Universitat Autònoma de Barcelona, which focuses on transferring academic knowledge to industry and society to address real-world communication and accessibility challenges.


The session featured six main presenters, each representing different aspects of accessibility research and implementation. Due to flight cancellations, some speakers participated online, creating a hybrid format. Speakers were given six-minute time slots to present their research projects efficiently before engaging in interactive discussions during the question-and-answer period.


## Major Presentations and Research Projects


### Easy-to-Understand News: The ENACT Project


Professor Ana Matamala from Universitat Autònoma de Barcelona presented the ENACT project, funded under the CREA scheme, which develops easy-to-understand news content across web, radio, and television platforms. The project addresses the principle that information should be accessible to all, with news serving as a crucial component of democratic participation.


The research revealed that easy-to-understand language exists on a continuum from easy language to plain language, addressing different user needs across diverse audiences including persons with intellectual disabilities, migrants, language learners, and others who may benefit from simplified news formats. The project conducted a comprehensive survey of European broadcasters, receiving 39 valid responses from 20 countries, finding that 13 organisations are already producing some form of easy-to-understand news content, primarily focusing on written website content.


The project has identified challenges in balancing accessibility requirements with journalistic principles. Professor Matamala noted that making news accessible whilst maintaining journalistic integrity requires careful consideration of editorial standards and professional practices. The project team plans to publish detailed survey results and reports on their website.


### AI-Enhanced Media Accessibility: Multiple Project Approach


Professor Pilar Orero presented three interconnected AI-enhanced projects addressing different aspects of media accessibility and bias. The Mosaic project creates a searchable hub for European public broadcaster archives, establishing a central processing system that enables users to search across multiple broadcaster collections simultaneously.


The Alfie project specifically addresses biases in media and AI models, particularly focusing on cultural and linguistic biases prevalent in European contexts. Professor Orero acknowledged the persistent nature of bias in AI systems, noting that “you will never eliminate bias” and that “AI reflects who we are, reflects humans, and humans are biased.”


Additionally, Professor Orero discussed a climate change communication project that addresses the accessibility gap in scientific data presentation, including making weather apps more accessible. She noted that whilst scientists produce excellent data, citizens often cannot access or understand this information due to presentation barriers.


### Virtual Reality in Education: The Inclusivity Project


Estella Oncins presented the Inclusivity project, an Erasmus Plus initiative exploring virtual reality technologies in educational contexts, particularly for neurodiverse students. The project aims to empower teachers and foster inclusive educational environments by addressing key challenges in social communication, teamwork, and problem-solving skills.


Initial pilot results demonstrated positive reception from both teachers and students, with particular enthusiasm for hands-on and creative activities. However, the research identified several implementation considerations, including time constraints (standard 1.5-hour allocations were insufficient for VR co-creation activities), easy language requirements, and the need for comprehensive storyboard templates for storytelling activities.


Future developments include creating a VR platform with embedded co-created social studies content and developing comprehensive implementation manuals for educators.


### Real-Time Transcription and Translation: The Scribal Project


Dr. Mireia Farrus from the University of Barcelona introduced Scribal, a real-time transcription and translation system specifically designed for university teaching environments. The system addresses accessibility needs for students with insufficient Catalan language command or hearing impairments by providing real-time transcription from Catalan to other languages.


The Scribal system utilises AI technology based on Whisper models, fine-tuned with specific databases including Common Voice and Parla en Parla for Catalan dialects and academic terminology. Importantly, the system is self-managed by universities to ensure privacy and security of data and content, addressing concerns about external data processing.


The project demonstrates particular attention to minority language preservation and support, specifically addressing the needs of Catalan speakers and learners in academic contexts. Dr. Farrus noted that the system will be presented at the Speech conference in Rotterdam for wider dissemination.


### Digital Accessibility as Educational Justice


Dr. Mar Gutierrez Colon from Universitat Rovira i Virgili presented two significant accessibility projects that frame digital inclusion as a matter of educational justice. Her work encompasses both international development and special educational needs contexts.


The first project involves the Tosomena Tuchese app, which addresses reading comprehension in Kiswahili for Kenyan children who primarily speak tribal languages. This gamified reading application serves underserved communities, addressing the fact that 9 out of 10 children in Sub-Saharan Africa cannot understand simple text. The application demonstrates how digital tools can bridge language gaps in educational contexts.


The second project focuses on developing English proficiency tests specifically designed for students with special educational needs and disabilities (SEND). This work enables these students to access labour market opportunities by providing appropriate assessment tools that accommodate their specific requirements.


Dr. Gutierrez Colon emphasised that digital accessibility represents educational justice, requiring design approaches that work with learners rather than for them. She noted that 240 million children globally have disabilities, and digital inclusion must enable effective learning beyond mere access provision.


## Key Themes and Insights


### Accessibility as Universal Design


The presenters consistently demonstrated that accessibility solutions benefit diverse audiences beyond traditional disability categories. Their tools serve multiple user groups including people with disabilities, migrants, language learners, and those with literacy challenges, emphasising the universal benefits of inclusive design approaches.


### AI as an Accessibility Tool


All presenters agreed that AI technology serves as a powerful tool for improving accessibility across different domains, from media content to educational materials. However, they also acknowledged the limitations and challenges, particularly regarding bias in AI systems.


### Participatory Design Approaches


Several speakers emphasised the importance of involving target users in the design process. Dr. Gutierrez Colon’s principle of designing “with, not for learners” represented a fundamental shift toward participatory design that includes target users as co-creators rather than passive recipients.


## Challenges and Questions Raised


### Training and Implementation Challenges


During the Q&A session, Professor NK Goel from India, representing connections to 100,000 colleges and universities, raised critical questions about training educators in rapidly evolving AI and quantum technologies. He noted that qualified experts are concentrated in companies that don’t typically cooperate with educational institutions, and mentioned having to drop AI courses due to lack of qualified teachers.


### Persistent Bias in AI Systems


Audience members expressed concerns about the persistence of bias in AI systems, particularly in translation systems used for fact-checking. While the Alfie project addresses these issues through research, the fundamental challenge of eliminating human bias from AI systems remains ongoing.


### Balancing Principles with Accessibility


The ENACT project highlighted the ongoing challenge of balancing journalistic principles with accessibility requirements in news production, an issue that requires further exploration and guideline development.


## Future Directions and Dissemination


Several concrete action items emerged from the presentations. The ENACT project will publish comprehensive survey results and reports on their website. The Inclusivity project plans to develop a guidebook for teachers implementing VR technologies and create a VR platform with embedded co-created content.


The Scribal system will be presented at international conferences for wider dissemination, whilst the Kenyan reading app research results will be published to inform similar initiatives globally. These dissemination efforts demonstrate commitment to knowledge transfer and scaling successful approaches.


## Conclusion


This discussion demonstrated how female academic leaders are successfully bridging university research with real-world societal problems through AI-enhanced accessibility solutions. The session revealed both the significant potential of current research and the ongoing challenges in implementation and scaling.


The presentations showed a mature field with coherent approaches and realistic expectations about both the possibilities and limitations of AI-enhanced accessibility solutions. The integration of technical innovation with social justice frameworks provides a strong foundation for continued development and implementation of accessibility technologies that serve diverse communities and promote inclusive access to information and education.


Session transcript

Anita Lamprecht: and gentlemen, I’m very glad to say welcome to all of you to our very late, but very important session about media and education for all, bridging female academic leaders in society towards impactful results. It’s my honor to moderate this session for you. My name is Dr. Anita Lamprecht. I come from the Diplo Foundation and I’m an emerging tech expert and also work as a researcher. But I want to thank especially the AXS network, CAT network today of the Universitat Autonoma de Barcelona for organizing this session. And as the session will demonstrate, the AXS-CAT is a knowledge transfer network that aims to transfer knowledge to industry and society. Well, I’d also like to extend a very warm welcome to our distinguished speakers. I have here on my right side, Estel la Oncinss and on my left side, Professor Pilar Orero. And online, we have Professor Ana Matamala, Dr. Mireia Farrus and Dr. Mar Gutierrez Colon. And thank you very much for joining us online. Our speakers unfortunately had their flights canceled, but thanks to technology, you can still participate in the session. Thank you. Yeah, our session today is about media and education for all. So we will learn how female experts bring ideas from the university to the lives of real people. So it’s really about solving real problems with the help of AI. And while the speakers will cover different projects, they all solve the core, in the core, the same problem. So the fact that some people cannot understand communication in the standard format is a significant issue. For them, it’s some kind of noise, it’s chaotic and without useful information, basically. So a standard news broadcast, for example, is clear to an average person, but it’s effectively useless noise for someone with a lower literacy skill or like a migrant who is confronted with a new language. And here, I’m very much looking forward to the presentation of Professor Anna Matamala. She will show us how to enhance standard news reports into clear and easy to understand information with the help of AI. Subtitles are another great way to make information understandable, especially when you speak a minority language. And here, Dr. Mireia Farrus will present us how AI can be useful as a tool for creating subtitles. Well, speaking English, we all know, is a game changer when it comes to access to information, but learning a language isn’t easy for everyone, especially for people with intellectual disability. So Dr. MarGutiérrez-Colón shows us how to transform a disability into a learning ability. Even for the average person, even for us experts, it can be very difficult and noisy when there is too much information and when the information has the wrong format. And here, Professor Pilar Orero will give us insights into the dedicated Mosaic project. Eventually, one of my personal favorites is virtuality. Dr. Estel la Oncins explores with us how the classroom can be extended into the virtual world to enhance students’ learning experience with a project. And here, you have to help me. How do you pronounce the project?


Estel la Oncins: Inclusivity. Inclusivity. It wasn’t easy.


Anita Lamprecht: That’s true. Excellent. So without further ado, we will now look how the speakers solve these problems that I showed you. Just a very quick organizational note. I will try to keep us on time. So I will give the speaker a notification when they have spoken for about five minutes so they can conclude the presentation after six minutes. And hopefully, we will have some time left so that you have also time to ask us some questions. So without further ado, Professor Ana Matamala from the Thomas University of Barcelona, the floor is yours. Thank you.


Ana Matamala: Thank you very much. I’m sorry that we cannot be there. Thank you very much for the introduction. In less than six, seven minutes, I will try to share with you, give you an overview of the ENACT project. So I will explain what the ENACT project is, the first steps we’ve taken in this project, and some of the challenges we have encountered. The ENACT project is a project about easy-to-understand news. On screen, the logo of the ENACT project, easy-to-understand news for collaborative transformation. It is funded under the CREA project, and the CREA, sorry, scheme. And the aim of the ENACT project is first, to explore the development of easy-to-understand language news in three main settings. So the web, the radio, and television. Of course, to achieve a more diverse media environment and increase media literacy among diverse audiences. The project is led by RTV Slovenia. We have broadcasters involved, RTV Slovenia, ORF, Latvias Radio, and the Catalan broadcaster CCMA. We are a university who’s taking part in that project, and there is also an association of persons with intellectual disabilities, UNIAMOCI. Just a short clarification of what we mean by easy-to-understand language. We think that this is a continuum in which of, or a continuum of language varieties which enhance comprehensibility. So we would have easy language, what was traditionally known as easy to read. And at the end of the continuum, we would have plain language. So our project covers all this spectrum, and we are aware that this implies differences in terms of language simplifications. And we also understand that we are addressing different uses. Why this project? Basically, this is very much aligned with SDG number 10. We think that information should be for all, and news are a fundamental part of this information. And they are also a very fundamental part of participation. So there is an interest in media to reach new audiences, to go beyond traditional access services. And when we started this project, we realised that some broadcasters had already started to produce some, let’s call them easy news. So what are we gonna do with this project? It’s a two-year project. During the first year, which is now exactly, what we’re doing is mapping and exchanging practices. So we’ve done a survey. I will summarise it very briefly in a few seconds. And we’re also doing some training and exchanges between partners so that we learn from what others are doing. And then next year, what we’ll do is we’ll produce some examples, taking inspiration from how others are producing easy to understand news. We will gather user feedback. And at the end, we will co-design a hub where we will feature some of the best practices we have found. As I told you, the first year has been knowing what is out there. So we produce a survey, the ANAC survey, which was shared with European broadcasters and organisations. We got 39 valid responses from 20 different countries. And we found out that from those who responded, 13 are producing some sort of easy to understand news. And also those include nine associations, but they’re mainly focused on written content on the website. We’ve seen that there are different approaches. Some focus more on easy. Some say that they produce some content in plain language. There is reference to persons with intellectual disabilities as a target audience, but also there is reference to migrants, to learners, to all people who may benefit from this easy to understand news format. And well, I think I don’t have time to get into much detail. I wanted to keep it very short. So if you are interested in finding out more about the survey percentages and so on, we will publish the information, the report on our website. On screen, I put a series of snapshots of some examples of easy to understand. Latvia’s radio is producing some easy radio content. RTV Slovenia have a portal, bottom right on screen, in which they include what they call easy to understand news, so it’s a website content. On the left, ORF, they offer Nachrichten, so news in plain language, and they have different videos and information. And on screen also a snapshot of a show, a news programme, that was presented and prepared in collaboration with an association of persons with Down syndrome. So those are some of the examples we’ve gathered. I don’t have time to get into it, but just to point out some of the challenges we found out in the first steps of this project. There’s been discussion about how to interact, how the interaction between accessibility and journalistic principles can take place. We’ve realised that there is still a lack of awareness and the need for further training. There is a need for clear guidelines, and we’ve realised that it is important that this becomes hopefully a strategic priority, because it’s still not a priority in many broadcasters. And with that, I will conclude my presentation. I will thank you for listening to me, and I understand that at the end of the session we’ll be able to take questions. If not, you can email me at anna.matamala.uab.gat. Thank you very much.


Anita Lamprecht: Thank you very much, Anna. I’ll pass the floor to Professor Pilar Orero.


Pilar Orero: Thank you very much. My presentation today… Please change the presentation. Pilar, it’s the first presentation. My presentation before, so I go on. No, no. No, it’s here. I am here. The next two presentations are people on site. Fantastic. Thank you very much. Excuse me. It’s not clicking the clicker. No. No. No. I’ll tell you, I’ll tell you and then you… Yeah? Now it’s going. Now it’s happening? Yes, it’s happening. Thank you very much. My presentation is called Enriching Media Accessibility with AI. The idea here is that the research that we’re doing right now, and this research paid by the European Commission, is to do with AI and how AI can really enhance accessibility services. And accessibility services are subtitling, audio description, sign language, easy to understand or easy to read, that sort of accessibility services that we have. The first course is not working now. The first project is that I would like to talk about is called Mosaic. Mosaic is to do with we are designing a single central processing hub for media content creators, for distributors and consumers. And the idea is that anybody would be able to go into that hub, that hub would end up being a search engine where you can really search in any European language, both by text or by voice. And you would be able to search in the archives of any public broadcaster in Europe. And then you can retrieve that information. And then you would be able to… Then the broadcasters that they don’t do very much with their archives will be able to sell that content or distribute the content further, because after all, it is public content made by, paid by public funding. So that is the first idea of the Mosaic. How do we search in the media that is already in the archives? We search through subtitling, which is one of the access services. So it is through subtitling that we can then understand what is the content of the media, we rescue it, we understand what is about it, and we do that. The next project that I would like to talk about is called Alfie. And Alfie, what we’re doing here is to understand what are the biases that we find in media and how that can be addressed. So the first thing that we have to understand is what, where do we find the biases? Everybody knows about the biases about the black faces, or know this, but there are many more biases beyond black faces. And they are, it is very interesting to point them out. Some of them are cultural, which in Europe, it is fantastic because we’re a very super rich continent. So we have to also take into consideration the languages and the cultures in Europe to understand the bias that we have when creating models for generative AI. The next, and I would really, really be very thankful if you could scan that and reply to that survey, please. The idea is we’re trying to understand what do you think is bias in data and how this should be processed. Thank you very much. And the final project that I’m going to present is to do with climate change, which is, again, very, very strange that I, who work for accessibility services, present a climate change project. The thing is that climate change presents data in a very unaccessible way. For example, the colors, the heat maps are not accessible. If you open your weather application in your mobile phone, you will probably understand only half of the information that is given to you. So what we are doing is trying to, we are working in this project with people who work in weather, scientists in weather, in climate change. And we are working together so they understand how to provide easier information for everyone to understand about the climate change. Because one of the problems is that scientists are producing fantastic data, but the citizen is not getting that information. So how we go across it, that is the idea. And this is my final presentation. Thank you very much.


Anita Lamprecht: Thank you very much, Professor Pilar Orero for this fantastic presentation and truly a showcase of how knowledge transfer can be taking place to the society. So our next speaker is Dr. Estel la Oncins, also from the Autonomous University of Barcelona. The floor is yours.


Estel la Oncins: Thank you very much. Thank you very much. I’ll wait until the presentation is on the screen. Yes. Perfect. Thank you very much. Thank you very much for the presentation. Now I’m going to jump into the virtual reality world through a project, an Erasmus Plus project that we are conducting and we are leading at Universitat Autònoma de Barcelona. It’s on exploring the use of virtual reality in class. And I will present some initial results from the inclusivity project. A short overview, just a few mentions on questions related to immersive environments, XR, in education. Then I will present the inclusivity project. It’s also nice and it’s also helpful in terms of inclusion and capacity building. But the main problem, and this is the sad face, is that there are major barriers that teachers and students face, such as the lack of accessible content, the lack of technical skills, and the lack of training and resources. So this is why we created the Inclusiviarity Project. It’s on VR technologies in teaching context. The three main objectives are empowering teachers with VR technologies, foster inclusive and accessible educational environments, especially with a focus on neurodiverse students, and develop skills such as social communication, teamwork, and problem solving, which according to first literature review that we’ve conducted, are main problems that neurodiverse students face, especially in educational context. Who are the partners? Universities, so we are the partners. We are the leaders of this project. And then, of course, if it’s in a teaching scenario, then we needed schools. We have a network of schools from Piera in Greece, a network of schools specialized with people with disabilities in Italy, and we also have the EMAP, which is the audio-visual school, with trainers based in Spain. Then we have an NGO in Cyprus, Citizens in Power, and, of course, if we are implementing a VR solution, then we needed a technical partner, and this is Ludisius in France. So we are an international consortium. Why? As mentioned, there are already existing VR solutions, but they are mainly standalone solutions, platforms, and they are not integrated in curricula. So what we are approaching is a sustainable development goal 4 and 10, so quality in education and reduced inequalities, and also what we are considering to match skills and include VR as part of the curricula. We are departing from the DIGICOM Framework 2.2, which is digital skills for all, not only for teachers but also for students. I’m not going to enter into what DIGICOM is, but it’s a European framework based on five dimensions, problem-solving, information and data literacy, communication and collaboration, digital content creation and safety. And we are also departing from the EU key competencies for lifelong learning, so based on the eight main, literacy, multilingual, mathematical slash science, digital, personal slash social, citizenship, entrepreneurship, and cultural awareness, because we want to include virtual reality for real in a curriculum. The first steps that we’ve conducted so far, we’ve started beginning of this year, so half a year more or less. We’ve already conducted a survey, 19 participants across Europe, all teachers and caregivers for neurodiverse students. We’ve conducted six focus groups in five countries, Spain, Italy, Greece, and Cyprus. These are mainly the countries with school networks, and we made the first pilot workshop in Spain. Initial results from a teacher’s perspective, these are the outcomes from the survey and the focus group, potential of VR use in education, especially to address the needs of neurodiverse students, potential of VR to use education to improve motivation. There is a clear need on receiving comprehensive training tools and to integrate this in educational practices. There is a recognized versatility of immersive technology in addressing all key competencies for lifelong learning. This was already reported in the survey and expressed during the focus groups, and the use of technologies and tools in teaching scenarios is considered very positive, especially in terms of hands-on and creative tasks. Initial results from students’ perspectives, this departs from the workshop that we’ve conducted with 16 students. The use of VR in methods in teaching scenarios is considered very positive, so all students reported to be very interested in having virtual reality in their teaching context. Hands-on activities and creative tasks are the most valued. This is also in line with teacher’s perspectives. Easy language is considered a need and a solution. Important also to consider questions related to time, especially in terms of co-creating sessions. We had a one hour and a half session. This was not enough, so when you are including VR in co-creating sessions with students, you have to take into consideration how long will this activity take. Then you also need to provide storyboard templates to support storytelling activities. And with all this, I’ll finish with the future steps. We will proceed to create a guidebook for teachers. We are also implementing a VR technology, a VR platform, with embedded co-created social studies, and an implementation manual and a scientific report on VR to address neurodiversity in education will be also implemented and developed. And this is all. Thank you. If you want to follow us, this is our webpage. Thank you very much for listening.


Anita Lamprecht: Thank you very much for this brilliant presentation. This actually sounds like a very futuristic approach and it sounds like fun. So our next presenter will present from online, please. So it’s Dr. Mireia Farrus from the University of Barcelona.


Mireia Farrus: Good afternoon. I can still not share the screen because there’s another one sharing. Yes, we don’t see on the screen yet. No, because I cannot share it. Estela, can you stop sharing? It’s the technical detail. Okay. Two seconds. Yeah, now. So, can you see? Can you see the slides? Yes, we can see. So good afternoon, everyone. Thank you very much for the invitation to this meeting. So I’m going to present Scribal. Scribal is a project for providing audio transcriptions in a university teaching environment. So… So Scribal stands for Digital Scribal. It is a real-time transcription system and translation system that aims at providing machine translation from Catalan language into other main languages spoken at universities, such Spanish, English, or other convenient languages. And one of the major strengths of this tool is that it is self-managed by our university, ensuring privacy and security of data and content. So Scribal is a tool that can be used from any device, any device from the student or any device in the classroom. And it covers several sustainable development goals, such as ensuring inclusive and equitable quality education, also increasing access to information and communications technology, or protecting the world’s cultural and natural heritage, the language in this case, especially when we speak about minority and minorised languages. So why do we need Scribal? This is the current situation. There are nowadays several existing automatic speech recognition and machine translation systems that do transcriptions and translations, but they are not trained specifically for an academic environment. Also, Scribal is a tool especially useful for those students without sufficient command of Catalan or students with hearing impairments. so that they can access to the transcriptions of the oral content once this is provided. Escriba is adapted to academic and scientific terminology and also adapted not only to Catalan language but also to different Catalan dialects because we have professors and lecturers from all around the country. So improving the accessibility and empowerment of the people at higher education. So due to a flight cancellation I am not able to be here with this today but from tomorrow I will be able to show you the Escriba interface and the tool which has both a presentation and listener roles for the lecturer and the student respectively. So the presenter can insert a code for the classroom and then if you want to access those classes as a listener you just type the code provided and you can choose whether you wish the transcription or the translation into your own language or any specific language. So Escriba is based on AI technology grounded on the Whisper open source models developed by the OpenAI company and what we do is to better tune the system to Catalan dialects using specific databases such as Common Voice and Parla en Parla which Common Voice is a universal oral speech database, Parla en Parla is specifically for Catalan. And we also work on the teaching plans of each subject provided by the same university to adapt the system to the university terminology. So we are now performing pilot tests at the University of Barcelona however the system can be scalable to other countries and universities especially those with minority languages facing the same problems as Catalan at higher education. Also the key facts reported by the World Health Organization show that a system for deaf people is more than useful nowadays we can see the numbers of people with hearing impairments due to the high rate of people with deafness and hearing loss. This is our wonderful team consisting of a wide variety of specialists like linguists, engineers, mathematicians, physicists and support staff that contribute to develop this pilot test and to the implementation of this tool at the university. And just to finish, Scribal will be presented at the speech conference in Rotterdam next August if you want to be there, we will present a show and demo there and I encourage you to visit our Stanford tomorrow to test it and discuss its potential for your own universities and countries and languages. Thank you very much for your attention.


Anita Lamprecht: Thank you very much for this great presentation and I’m sure this can also help to foster the exchange of students around the world with such opportunities. The next speaker is Dr. Mar Gutierrez Colon from Universitat Rovira Virgili I give you the floor, thank you very much.


Mar Gutierrez Colon: Thank you very much. Thank you, I had my microphone off. So, my presentation is going to be based on two studies that my research group is conducting. Since I have very few time and I’m the last one, I’m going to try to be very quick but you have my email at the end in case you want to ask me some questions, right? So, the accessibility challenge that we propose in these two projects are based on these general accessibility challenges, right? 240 million children globally have disabilities, yes? And digital inclusion must go beyond access. It must enable effective learning. As my colleagues have said, we have many, many different tools but probably many of them are not accessible to these special needs students. So, our first project, the one that I’m going to present you is called Tosomena Tuchese. It’s in Kiswahili. It means let’s read and play. So, what happens in Kenya is that kids… So, there are two official languages, English and Swahili and many kids in rural areas only speak their tribal language and they have to learn Kiswahili, sorry, Swahili to enter school and there are many difficulties finding materials, not to say technology. So, what we have done is we have created an app, a gamified app to improve reading comprehension in Kiswahili, yeah? We have taken into account that these children may not be used to using technology, may not have or be used to use mobile phones, there is no Wi-Fi, well, there is not many things in those schools. So, the first thing we had to do is teach them how to use that, help students with the technology and then use the app. The figures are well-known and shocking. In Sub-Saharan Africa, 9 out of 10 children cannot understand a simple text. And girls, the numbers in girls is even worse. This is why we started this project, right? So, Tsomen Apites has large fonts, audio narration, intuitive gestures, is adapted for emerging readers. We have to imagine that they are very young. It is ideal for underserved communities and we have tested the app in real classrooms. In fact, I spent three months and then two months in Nairobi and travelled around different schools. So, we have the final results that are going to be published soon. We visited more than 30 schools and we tried the app. We have a new project with an improved version of the app. Basically, this is what we have. The second project is an English proficiency test for SEND students. And this is based on the necessity of special needs students to have a level test to assess, for them to get a certificate on their level of English so they can access to the labour market. We tried to get those certificates through the usual language test that we have but it is quite difficult for them to be able to answer all the questions, to follow the text, to read all the text, because they are very long, etc. So, what we are doing is working with special schools and trying to develop this exam. This test will provide multiple means of engagement like games, visuals, audio. Support multiple means of expression, speech, gestures, touch. And then it empowers both learners and educators because they’ve told us that when they teach English, basic English to these students, they find that they have an objective that is past this test in order to go to the labour market. And it enables multiple means of representation, multi-lingual, multi-modal content as it should be. When it has research, research goes to policy impact. We are embedded in AccessCAD mission and I think that we have very good collaboration in terms of we have moved our language acquisition research into this accessibility language acquisition goal in our research group. We are trying to find funding from different governments to be able to include all these or to spread all these projects to different countries. So, we believe that digital accessibility is educational justice. and we must design with, not for learners, yes. Accessibility strength outcomes for everyone and it bridges innovation and equity, as you know. We don’t have much time but my email is here and thank you very much for listening.


Anita Lamprecht: Thank you very much, Dr. Mar. As a lawyer, I found this especially impressive that you said digital accessibility is a form of educational justice. So I think that’s a very good statement. So thank you very much. All the speakers were very fast and precise in the timing. So we still have very valuable five minutes for our guests here in the room and online. I’m sorry, you see even faster than my timer. So we would be very much looking to your questions. It’s a very good opportunity to meet five distinguished experts in this very, very important field. Yes, please.


Audience: NK Goel from India. I came late here. Hello. Yeah, I am Professor NK Goel from India. I have not been able to listen to what you said in the beginning. But my problem is that lots of things being said on the new technologies, AI, quantum, etc. How do we teach them in the university colleges because we don’t get the professors or trained people and technology developers of the world, they do not teach. So we just live on the Google or chat. What is the solution we have?


Pilar Orero: I don’t know in the other universities, but in my university, there is the possibility. In fact, it’s not a possibility. What we do is that we do have some teachers that they are from industry, not academics. They have six hours of teaching only. And we get them for medicine. We get them for all these subjects where media, the industry is running faster than academics. And it is we rely on them to come in and to bring the point of view of industry into our students. And these teachers usually for them is very good to do this teaching and for the students is very good to have them. Also, because they are then able to understand what’s happening in industry and perhaps do internships with them. That’s how we do it in Catalonia, in where we come from. I don’t know in your country if you can do that. You have the possibility.


Audience: Actually, we do have this system because in brief, I am connected to 100,000 colleges, universities in India. And the problem is in the case of traditional industries like cement, steel, chemical, paper, we have this facility. But in the case of new technologies like AI, quantum, all these things, the experts are limited from two, three companies. We know the names. So they don’t cooperate. That is a problem we have. In fact, I will just take half a minute. We started as a compulsory course for AI and data for students in class 10th and 12th. For one year, we could not get a teacher, so we dropped the program, officially in the government.


Pilar Orero: Yes, you’re absolutely right. I mean, if we have to depend on the big names, forget it.


Anita Lamprecht: Do we have another question in the room? Yes, please.


Audience: Hi, everyone. My name is Tamiris. I have a question to you, Professor. We work with fact-checking, and we know the importance of using the right word to refer a fact. So I was thinking about the forms that you shared with us, and I was thinking about the bias in the process of translation. Because what is your experience with this, and what have you found? And what is your point of view about the bias in this process?


Pilar Orero: Thank you very much for your question. It’s terrible, because it is true. And also because you will never eliminate bias, whatever you want, I mean, whatever you’ve done. Yes, in the languages, there is the very, like in faces, you know, AI and black faces. In translation, we have the issue with the gender. I mean, feminine and masculine. That is always there. It’s very stupid, but it’s always wrong, and it’s always there, and it’s persistent. And even in really good Google translation, it’s persistent there. There are other more subtle issues as well, that I don’t think we’re going to never be able to get rid of them. And I think also because AI reflects who we are, reflects humans, and humans are biased, and are biased in the language that they use. So, yeah, I don’t know. Anna Matamala, can you have a better answer than mine?


Ana Matamala: No, I don’t have a better answer. I think it’s pretty good.


Anita Lamprecht: Yeah, excellent. Thank you so much. Well, AI is indeed about probability, so we have the bias, so very well said. And when it comes to AI education, so I think it’s us educators who have to upgrade our knowledge to support the students. So the Diplo Foundation is, for example, doing this with AI apprenticeship, so you’re very welcome also to explore. Yes, please.


Audience: I agree with you. We, the educators, have to upscale. My problem is when you teach MBA, MTech, BTech, there is a course, there is a syllabus, there is a system. When you have to teach new technologies, there is no course, no syllabus, nobody approves it. And it is a real problem I’m facing because every university asks me for the professor or teacher, and we don’t get it. So the problem is upscale is fine, but who will upscale you? And with the things going on in the ChatJPT, you know, all the three, four companies, DeepSeek, et cetera, everything is new coming someday. How do you train the teachers? Only thing is the internet website. And now ChatJPT, I don’t want to comment. You can see the experience of what they do. It is said that all the data of ChatJPT is prior to three, four years. After that, data has been scrambled. It’s a real problem.


Anita Lamprecht: Does one of our speakers want to respond to that? Ana, you’re the organizer of this session together with the Access Cats Network. So I want to give you the honor of the closing words.


Ana Matamala: Thank you. Thank you very much, Anit. I think that the comments that have been made are very relevant and point at one direction, which is the collaboration of industry and academia. And that happens at different levels. It can happen when we design or co-design sessions, educational programs, as the colleague from the audience was mentioning, that allow us to reskill and upskill. But also when we try to transfer the knowledge we generate in research projects as the one that have been presented into society, into industry. So to sum it up in 10 seconds, well, 20, I think that what we’ve tried in this session is to showcase some projects of research groups in the Access Cat Network. And Access Cat, what we are trying to do is support researchers so that this knowledge transfer, this knowledge valorization happens. So again, we apologize for not being there. And we are really grateful from Access Cat to all the speakers and the fantastic moderator that we could have in this session. Thank you very much.


Anita Lamprecht: Yeah, thank you. With these words, we close the session. And I must say, it is very impressive to have five female experts on this topic in one room, even if we share it virtually and physically. So we’re actually in a metaverse already. Thank you so much. And goodbye. Thank you. Bye bye. Maybe Mireia-Marie, if you can stay for a second, please. Sure.


P

Pilar Orero

Speech speed

124 words per minute

Speech length

1023 words

Speech time

494 seconds

AI can enhance accessibility services like subtitling, audio description, and sign language interpretation

Explanation

Professor Orero argues that AI technology can significantly improve existing accessibility services that help make media content accessible to people with disabilities. This represents a technological advancement in making information more inclusive.


Evidence

The research is funded by the European Commission and focuses on how AI can enhance accessibility services including subtitling, audio description, sign language, and easy to understand content.


Major discussion point

AI enhancement of accessibility services


Topics

Rights of persons with disabilities | Online education | Digital access


Agreed with

– Ana Matamala
– Mireia Farrus
– Mar Gutierrez Colon

Agreed on

AI technology can enhance accessibility services and make information more inclusive


The Mosaic project creates a central processing hub for media content that enables searching across European public broadcaster archives

Explanation

The Mosaic project aims to create a unified search engine where users can search in any European language, both by text or voice, across archives of European public broadcasters. This would allow better utilization and distribution of publicly funded content.


Evidence

The hub would function as a search engine accessible in any European language, allowing broadcasters to sell or distribute archived content that was originally paid for by public funding.


Major discussion point

Centralized media content accessibility


Topics

Digital access | Cultural diversity | Multilingualism


The Alfie project addresses biases in media and AI models, particularly cultural and linguistic biases in Europe

Explanation

Professor Orero explains that the Alfie project focuses on identifying and addressing various types of biases in media beyond the commonly known racial biases. The project recognizes Europe’s cultural and linguistic diversity as both a strength and a challenge for AI model development.


Evidence

The project includes a survey to understand public perception of bias in data and how it should be processed, acknowledging that biases extend beyond racial issues to include cultural and linguistic elements.


Major discussion point

Bias in AI and media systems


Topics

Cultural diversity | Multilingualism | Human rights principles


Disagreed with

– Audience

Disagreed on

Approach to addressing AI bias in translation systems


Climate change data is presented in inaccessible ways, requiring collaboration with scientists to make information more understandable for citizens

Explanation

Professor Orero argues that climate change information, such as heat maps and weather applications, is often presented in ways that are not accessible to the general public. The project involves collaboration between accessibility experts and climate scientists to bridge this communication gap.


Evidence

Examples include inaccessible colors in heat maps and weather applications where users can only understand half of the information provided, highlighting the disconnect between scientific data production and citizen comprehension.


Major discussion point

Accessibility of scientific information


Topics

Digital access | Rights of persons with disabilities | Sustainable development


A

Ana Matamala

Speech speed

145 words per minute

Speech length

1065 words

Speech time

437 seconds

The ENACT project develops easy-to-understand language news across web, radio, and television to increase media literacy among diverse audiences

Explanation

Professor Matamala presents the ENACT project as a two-year initiative that aims to create more accessible news content across multiple media platforms. The project seeks to achieve a more diverse media environment and improve media literacy for various audience groups.


Evidence

The project is funded under the CREA scheme and involves multiple European broadcasters including RTV Slovenia, ORF, Latvias Radio, and CCMA, along with universities and disability organizations.


Major discussion point

Accessible news content development


Topics

Digital access | Rights of persons with disabilities | Online education


Agreed with

– Pilar Orero
– Mireia Farrus
– Mar Gutierrez Colon

Agreed on

AI technology can enhance accessibility services and make information more inclusive


Easy-to-understand language exists on a continuum from easy language to plain language, addressing different user needs

Explanation

Professor Matamala explains that easy-to-understand language is not a single format but rather a spectrum of language varieties that enhance comprehensibility. This continuum ranges from traditional ‘easy to read’ formats to plain language, with different levels of simplification for different audiences.


Evidence

The project covers this entire spectrum and acknowledges that different points on the continuum require different approaches to language simplification and serve different user groups.


Major discussion point

Language accessibility spectrum


Topics

Rights of persons with disabilities | Multilingualism | Digital access


Survey results show 13 out of 39 European broadcasters are producing some form of easy-to-understand news, mainly focusing on written website content

Explanation

The ENACT survey of European broadcasters and organizations revealed that a significant portion are already producing accessible news content, though primarily in written format on websites. This indicates both existing progress and room for expansion into other media formats.


Evidence

The survey received 39 valid responses from 20 different countries, with 13 organizations producing easy-to-understand news, including 9 associations, mostly focused on written website content.


Major discussion point

Current state of accessible news production


Topics

Digital access | Rights of persons with disabilities | Freedom of the press


Target audiences include persons with intellectual disabilities, migrants, learners, and anyone who may benefit from simplified news formats

Explanation

Professor Matamala emphasizes that easy-to-understand news serves a broad range of audiences beyond just people with intellectual disabilities. The inclusive approach recognizes that many different groups can benefit from more accessible information formats.


Evidence

Survey findings showed references to multiple target groups including persons with intellectual disabilities, migrants, language learners, and the general population who may benefit from simplified news formats.


Major discussion point

Diverse audiences for accessible content


Topics

Rights of persons with disabilities | Digital access | Inclusive finance


Agreed with

– Mireia Farrus
– Mar Gutierrez Colon
– Anita Lamprecht

Agreed on

Accessibility solutions must address diverse audiences beyond just people with disabilities


E

Estel la Oncins

Speech speed

129 words per minute

Speech length

882 words

Speech time

408 seconds

The Inclusivity project explores VR technologies in teaching contexts to empower teachers and foster inclusive educational environments for neurodiverse students

Explanation

Dr. Oncins presents the Inclusivity project as an Erasmus Plus initiative that aims to integrate virtual reality into educational settings with a specific focus on supporting neurodiverse students. The project has three main objectives: empowering teachers with VR technologies, fostering inclusive environments, and developing key skills.


Evidence

The project involves an international consortium including universities, schools specialized in disabilities, NGOs, and technical partners across Spain, Greece, Italy, Cyprus, and France.


Major discussion point

VR integration in inclusive education


Topics

Online education | Rights of persons with disabilities | Digital access


VR can address skills development in social communication, teamwork, and problem-solving, which are main challenges for neurodiverse students

Explanation

Dr. Oncins argues that virtual reality technology can specifically target the development of crucial skills that neurodiverse students typically struggle with in educational contexts. The project is designed based on literature review identifying these key areas of difficulty.


Evidence

A literature review conducted by the project team identified social communication, teamwork, and problem-solving as main problems that neurodiverse students face, especially in educational contexts.


Major discussion point

Skill development through VR for neurodiverse students


Topics

Rights of persons with disabilities | Online education | Digital access


Initial results show both teachers and students view VR positively in educational settings, particularly for hands-on and creative tasks

Explanation

Dr. Oncins reports that preliminary findings from surveys, focus groups, and pilot workshops demonstrate positive reception of VR technology from both educators and students. The technology is particularly valued for interactive and creative learning activities.


Evidence

Results from 19 teacher participants across Europe, six focus groups in five countries, and a pilot workshop with 16 students all showed positive attitudes toward VR use in education, with particular enthusiasm for hands-on and creative tasks.


Major discussion point

Positive reception of VR in education


Topics

Online education | Digital access | Rights of persons with disabilities


Agreed with

– Mar Gutierrez Colon

Agreed on

Hands-on and creative approaches are most effective in educational technology implementation


Implementation requires consideration of time constraints, easy language needs, and provision of storyboard templates for storytelling activities

Explanation

Dr. Oncins identifies practical implementation challenges that emerged from pilot testing, including the need for adequate time allocation, simplified language use, and structured support materials. These findings inform best practices for VR integration in educational settings.


Evidence

The pilot workshop revealed that a 1.5-hour session was insufficient for co-creating activities, students identified easy language as a need and solution, and storyboard templates were necessary to support storytelling activities.


Major discussion point

Practical implementation considerations for VR in education


Topics

Online education | Rights of persons with disabilities | Digital access


M

Mireia Farrus

Speech speed

123 words per minute

Speech length

696 words

Speech time

338 seconds

The Scribal project provides real-time transcription and translation from Catalan to other languages in university teaching environments

Explanation

Dr. Farrus presents Scribal as a comprehensive solution for making university lectures accessible through real-time transcription and translation services. The system specifically addresses the needs of Catalan-language instruction while providing access to other major languages used in universities.


Evidence

Scribal provides machine translation from Catalan to Spanish, English, and other convenient languages, and can be used from any device in the classroom or by students.


Major discussion point

Real-time language accessibility in higher education


Topics

Multilingualism | Online education | Rights of persons with disabilities


The system is self-managed by universities to ensure privacy and security of data and content

Explanation

Dr. Farrus emphasizes that one of Scribal’s major strengths is its institutional control, allowing universities to maintain data privacy and security rather than relying on external commercial services. This addresses concerns about sensitive academic content and student information.


Evidence

The tool is self-managed by the university, ensuring privacy and security of data and content, distinguishing it from existing commercial automatic speech recognition and machine translation systems.


Major discussion point

Data privacy in educational technology


Topics

Privacy and data protection | Online education | Digital access


Scribal is adapted to academic terminology and different Catalan dialects, improving accessibility for students with insufficient Catalan command or hearing impairments

Explanation

Dr. Farrus explains that Scribal is specifically designed for the academic environment and linguistic diversity of Catalonia, addressing both terminological precision and dialectal variation. This makes it particularly useful for students who face language or hearing barriers.


Evidence

The system is adapted to academic and scientific terminology and different Catalan dialects because professors and lecturers come from all around the country, serving students without sufficient command of Catalan or students with hearing impairments.


Major discussion point

Specialized accessibility for academic contexts


Topics

Rights of persons with disabilities | Multilingualism | Online education


Agreed with

– Ana Matamala
– Mar Gutierrez Colon
– Anita Lamprecht

Agreed on

Accessibility solutions must address diverse audiences beyond just people with disabilities


The system uses AI technology based on Whisper models, fine-tuned with specific databases for Catalan dialects

Explanation

Dr. Farrus describes the technical foundation of Scribal, which builds upon existing open-source AI models but enhances them with specialized training data for Catalan language varieties. This approach combines proven technology with localized adaptation.


Evidence

Scribal is based on Whisper open source models developed by OpenAI, fine-tuned using specific databases such as Common Voice and Parla en Parla, with the latter being specifically for Catalan, plus teaching plans from the university to adapt to university terminology.


Major discussion point

AI adaptation for minority languages


Topics

Multilingualism | Online education | Cultural diversity


Agreed with

– Pilar Orero
– Ana Matamala
– Mar Gutierrez Colon

Agreed on

AI technology can enhance accessibility services and make information more inclusive


M

Mar Gutierrez Colon

Speech speed

132 words per minute

Speech length

722 words

Speech time

326 seconds

Digital accessibility represents educational justice, requiring design with learners rather than for learners

Explanation

Dr. Gutierrez Colon frames digital accessibility as a fundamental issue of educational equity and justice. She emphasizes the importance of participatory design approaches that involve learners in the development process rather than simply creating solutions for them without their input.


Evidence

The statement ‘digital accessibility is educational justice’ and the principle that ‘we must design with, not for learners’ represents the philosophical foundation of her research approach.


Major discussion point

Educational justice through digital accessibility


Topics

Rights of persons with disabilities | Digital access | Human rights principles


Agreed with

– Pilar Orero
– Ana Matamala
– Mireia Farrus

Agreed on

AI technology can enhance accessibility services and make information more inclusive


The Tosomena Tuchese app addresses reading comprehension in Kiswahili for Kenyan children who only speak tribal languages

Explanation

Dr. Gutierrez Colon describes a gamified app designed to help Kenyan children transition from their tribal languages to Kiswahili, which is required for school entry. The project addresses the linguistic and technological barriers faced by children in rural areas with limited resources.


Evidence

The app name means ‘let’s read and play’ in Kiswahili, features large fonts, audio narration, and intuitive gestures adapted for emerging readers. The research involved testing in over 30 schools with the researcher spending three months in Nairobi.


Major discussion point

Language learning technology for underserved communities


Topics

Multilingualism | Digital access | Online education


Agreed with

– Estel la Oncins

Agreed on

Hands-on and creative approaches are most effective in educational technology implementation


Development of English proficiency tests for students with special educational needs enables access to labor market opportunities

Explanation

Dr. Gutierrez Colon argues that specialized English proficiency testing is necessary for students with special educational needs to obtain certificates that enable labor market access. Traditional language tests are not suitable for these students due to their format and complexity.


Evidence

The project works with special schools to develop tests that provide multiple means of engagement through games, visuals, and audio, and support multiple means of expression including speech, gestures, and touch.


Major discussion point

Accessible language certification for employment


Topics

Rights of persons with disabilities | Future of work | Digital access


240 million children globally have disabilities, and digital inclusion must enable effective learning beyond mere access

Explanation

Dr. Gutierrez Colon presents the global scale of the accessibility challenge and argues that simply providing access to technology is insufficient. True digital inclusion requires ensuring that technology actually enables effective learning outcomes for children with disabilities.


Evidence

The statistic of 240 million children globally having disabilities provides the quantitative foundation for the argument, while the distinction between access and effective learning highlights the qualitative dimension of the challenge.


Major discussion point

Global scale of digital accessibility needs


Topics

Rights of persons with disabilities | Digital access | Sustainable development


Agreed with

– Ana Matamala
– Mireia Farrus
– Anita Lamprecht

Agreed on

Accessibility solutions must address diverse audiences beyond just people with disabilities


A

Audience

Speech speed

132 words per minute

Speech length

447 words

Speech time

203 seconds

Universities face difficulties finding qualified professors for new technologies like AI and quantum computing, as industry experts from major companies don’t typically teach

Explanation

An audience member from India, Professor NK Goel, argues that while universities can find industry professionals to teach traditional subjects, experts in cutting-edge technologies like AI and quantum computing are concentrated in a few major companies and are not available for teaching. This creates a significant gap in technology education.


Evidence

The speaker mentions being connected to 100,000 colleges and universities in India, and describes how traditional industries like cement, steel, and chemicals provide industry teachers, but AI and quantum experts from major companies don’t cooperate. India even had to drop a mandatory AI course for students because they couldn’t find teachers.


Major discussion point

Shortage of qualified technology educators


Topics

Online education | Future of work | Capacity development


Traditional collaboration between industry and academia works for established fields but fails for emerging technologies where expertise is concentrated in few companies

Explanation

The audience member explains that while industry-academia collaboration functions well for traditional industries, it breaks down for emerging technologies where expertise is limited to a small number of major technology companies. This concentration of knowledge creates barriers to educational access.


Evidence

The comparison between traditional industries (cement, steel, chemical, paper) where industry-academia collaboration works versus new technologies (AI, quantum) where experts are limited to ‘two, three companies’ and ‘don’t cooperate.’


Major discussion point

Limitations of industry-academia collaboration in emerging tech


Topics

Future of work | Capacity development | Digital business models


Bias in AI translation systems is persistent and reflects human biases, particularly in gender translation and other subtle linguistic issues

Explanation

An audience member working in fact-checking raises concerns about bias in AI translation systems, particularly regarding accurate word choice in fact-checking contexts. Professor Orero responds that bias in AI systems is persistent and reflects human biases, with gender translation being a particularly problematic area.


Evidence

Examples include persistent gender bias in translation systems, even in advanced systems like Google Translate, and the acknowledgment that AI reflects human biases because it’s trained on human-generated data.


Major discussion point

Persistent bias in AI language systems


Topics

Human rights principles | Multilingualism | Gender rights online


Disagreed with

– Pilar Orero

Disagreed on

Approach to addressing AI bias in translation systems


Knowledge transfer between academia and industry requires collaboration at multiple levels, including co-designed educational programs and research project implementation

Explanation

The moderator, Anita Lamprecht, synthesizes the discussion by arguing that effective knowledge transfer requires multi-level collaboration between academia and industry. This includes both the design of educational programs and the implementation of research projects in real-world settings.


Evidence

The session itself serves as evidence, showcasing research projects from the Access Cat Network that demonstrate knowledge transfer from university research to societal applications, with Ana Matamala noting that Access Cat supports researchers in knowledge valorization.


Major discussion point

Multi-level collaboration for knowledge transfer


Topics

Capacity development | Digital business models | Interdisciplinary approaches


A

Anita Lamprecht

Speech speed

140 words per minute

Speech length

1044 words

Speech time

445 seconds

Standard communication formats create barriers for people with lower literacy skills or migrants facing new languages, making information appear as useless noise

Explanation

Dr. Lamprecht argues that what appears as clear communication to average people becomes incomprehensible noise for those with literacy challenges or language barriers. She emphasizes that standard news broadcasts, while clear to most, are effectively useless for people with lower literacy skills or migrants confronting new languages.


Evidence

She provides the example of standard news broadcasts being clear to average persons but effectively useless noise for someone with lower literacy skills or migrants confronted with a new language.


Major discussion point

Communication barriers in standard media formats


Topics

Digital access | Rights of persons with disabilities | Multilingualism


Agreed with

– Ana Matamala
– Mireia Farrus
– Mar Gutierrez Colon

Agreed on

Accessibility solutions must address diverse audiences beyond just people with disabilities


Female academic leaders demonstrate how university research can solve real-world problems through AI applications

Explanation

Dr. Lamprecht frames the session as showcasing how female experts bring ideas from universities to real people’s lives, solving practical problems with AI technology. She emphasizes that while the speakers cover different projects, they all address the core problem of making communication accessible to those who cannot understand standard formats.


Evidence

The session features five female experts presenting projects that use AI to solve accessibility problems, from news simplification to subtitle creation to language learning for people with disabilities.


Major discussion point

Knowledge transfer from academia to society


Topics

Capacity development | Digital access | Gender rights online


AI education requires educators to upgrade their knowledge to support students, with institutions like Diplo Foundation providing AI apprenticeship programs

Explanation

Dr. Lamprecht responds to concerns about AI education by arguing that educators themselves must take responsibility for upgrading their knowledge to support students. She suggests that educational institutions are developing programs to address this need, citing the Diplo Foundation’s AI apprenticeship as an example.


Evidence

She mentions the Diplo Foundation’s AI apprenticeship program as an example of how educational institutions are addressing the need for AI education and invites participants to explore such opportunities.


Major discussion point

Educator responsibility in AI education


Topics

Capacity development | Online education | Future of work


The session demonstrates a metaverse-like collaboration by bringing together five female experts both virtually and physically

Explanation

Dr. Lamprecht observes that the session itself represents an innovative form of collaboration, with speakers participating both online and in-person due to flight cancellations. She notes this as impressive representation of female expertise in the field and suggests they are already operating in a metaverse-like environment.


Evidence

The session included speakers both online (due to flight cancellations) and physically present, with Dr. Lamprecht noting they were ‘sharing it virtually and physically’ and were ‘actually in a metaverse already.’


Major discussion point

Hybrid collaboration models


Topics

Online education | Gender rights online | Digital access


Agreements

Agreement points

AI technology can enhance accessibility services and make information more inclusive

Speakers

– Pilar Orero
– Ana Matamala
– Mireia Farrus
– Mar Gutierrez Colon

Arguments

AI can enhance accessibility services like subtitling, audio description, and sign language interpretation


The ENACT project develops easy-to-understand language news across web, radio, and television to increase media literacy among diverse audiences


The system uses AI technology based on Whisper models, fine-tuned with specific databases for Catalan dialects


Digital accessibility represents educational justice, requiring design with learners rather than for learners


Summary

All speakers agree that AI technology serves as a powerful tool for improving accessibility across different domains – from media content to educational materials – making information more inclusive for diverse audiences including people with disabilities, language learners, and underserved communities.


Topics

Digital access | Rights of persons with disabilities | Online education


Accessibility solutions must address diverse audiences beyond just people with disabilities

Speakers

– Ana Matamala
– Mireia Farrus
– Mar Gutierrez Colon
– Anita Lamprecht

Arguments

Target audiences include persons with intellectual disabilities, migrants, learners, and anyone who may benefit from simplified news formats


Scribal is adapted to academic terminology and different Catalan dialects, improving accessibility for students with insufficient Catalan command or hearing impairments


240 million children globally have disabilities, and digital inclusion must enable effective learning beyond mere access


Standard communication formats create barriers for people with lower literacy skills or migrants facing new languages, making information appear as useless noise


Summary

Speakers consistently recognize that accessibility solutions serve multiple user groups including people with disabilities, migrants, language learners, and those with literacy challenges, emphasizing the broad societal impact of inclusive design.


Topics

Rights of persons with disabilities | Digital access | Multilingualism


Hands-on and creative approaches are most effective in educational technology implementation

Speakers

– Estel la Oncins
– Mar Gutierrez Colon

Arguments

Initial results show both teachers and students view VR positively in educational settings, particularly for hands-on and creative tasks


The Tosomena Tuchese app addresses reading comprehension in Kiswahili for Kenyan children who only speak tribal languages


Summary

Both speakers found through their research that interactive, hands-on, and creative educational approaches are most valued and effective, whether in VR environments or gamified learning applications.


Topics

Online education | Digital access | Rights of persons with disabilities


Similar viewpoints

Both speakers emphasize the importance of institutional control and awareness of systemic issues in AI systems – Orero focusing on bias recognition and Farrus on data privacy and security in educational contexts.

Speakers

– Pilar Orero
– Mireia Farrus

Arguments

The system is self-managed by universities to ensure privacy and security of data and content


The Alfie project addresses biases in media and AI models, particularly cultural and linguistic biases in Europe


Topics

Privacy and data protection | Human rights principles | Online education


Both speakers recognize that accessibility is not one-size-fits-all but requires nuanced approaches with different levels of simplification and support materials tailored to specific user needs and contexts.

Speakers

– Ana Matamala
– Estel la Oncins

Arguments

Easy-to-understand language exists on a continuum from easy language to plain language, addressing different user needs


Implementation requires consideration of time constraints, easy language needs, and provision of storyboard templates for storytelling activities


Topics

Rights of persons with disabilities | Online education | Digital access


Both speakers frame accessibility work as a matter of justice and empowerment, emphasizing participatory approaches and the role of academic research in addressing real-world social problems.

Speakers

– Mar Gutierrez Colon
– Anita Lamprecht

Arguments

Digital accessibility represents educational justice, requiring design with learners rather than for learners


Female academic leaders demonstrate how university research can solve real-world problems through AI applications


Topics

Human rights principles | Gender rights online | Capacity development


Unexpected consensus

Bias in AI systems is persistent and reflects inherent human biases

Speakers

– Pilar Orero
– Audience

Arguments

The Alfie project addresses biases in media and AI models, particularly cultural and linguistic biases in Europe


Bias in AI translation systems is persistent and reflects human biases, particularly in gender translation and other subtle linguistic issues


Explanation

The unexpected consensus emerged when an audience member raised concerns about bias in AI translation for fact-checking, and Professor Orero candidly acknowledged that bias is persistent and unavoidable because AI reflects human biases. This honest admission of limitations in AI systems was surprising given the generally optimistic tone about AI applications throughout the session.


Topics

Human rights principles | Multilingualism | Gender rights online


Industry-academia collaboration has significant limitations for emerging technologies

Speakers

– Audience
– Pilar Orero
– Anita Lamprecht

Arguments

Universities face difficulties finding qualified professors for new technologies like AI and quantum computing, as industry experts from major companies don’t typically teach


Traditional collaboration between industry and academia works for established fields but fails for emerging technologies where expertise is concentrated in few companies


Knowledge transfer between academia and industry requires collaboration at multiple levels, including co-designed educational programs and research project implementation


Explanation

An unexpected consensus emerged around the challenges of technology education, with speakers acknowledging that traditional industry-academia collaboration models break down for cutting-edge technologies where expertise is concentrated in a few major companies. This was unexpected given the session’s focus on successful knowledge transfer projects.


Topics

Future of work | Capacity development | Digital business models


Overall assessment

Summary

The speakers demonstrated strong consensus on the fundamental principles of accessibility and inclusion, the potential of AI to enhance accessibility services, the need to serve diverse audiences beyond traditional disability categories, and the importance of participatory design approaches. They also agreed on practical implementation considerations such as the effectiveness of hands-on learning approaches.


Consensus level

High level of consensus with significant implications for the field. The agreement among these female academic leaders suggests a mature understanding of accessibility challenges and a shared vision for AI-enhanced solutions. The consensus extends beyond technical approaches to include philosophical frameworks about justice, inclusion, and participatory design. However, the honest acknowledgment of persistent challenges like AI bias and educational barriers demonstrates a realistic rather than overly optimistic perspective. This level of agreement among experts from different institutions and countries suggests that the field has developed coherent principles and approaches, which could facilitate broader adoption and implementation of accessibility technologies.


Differences

Different viewpoints

Approach to addressing AI bias in translation systems

Speakers

– Pilar Orero
– Audience

Arguments

The Alfie project addresses biases in media and AI models, particularly cultural and linguistic biases in Europe


Bias in AI translation systems is persistent and reflects human biases, particularly in gender translation and other subtle linguistic issues


Summary

While both acknowledge the existence of bias in AI systems, Professor Orero presents an active research approach to identify and address biases through the Alfie project, while the audience member expresses more pessimism about the persistence of bias, particularly in translation systems. Orero suggests biases can be studied and potentially mitigated, while the audience perspective implies they may be inherently persistent.


Topics

Human rights principles | Multilingualism | Cultural diversity


Unexpected differences

Feasibility of eliminating bias in AI systems

Speakers

– Pilar Orero
– Audience

Arguments

The Alfie project addresses biases in media and AI models, particularly cultural and linguistic biases in Europe


Bias in AI translation systems is persistent and reflects human biases, particularly in gender translation and other subtle linguistic issues


Explanation

This disagreement was unexpected because it emerged from an audience question about fact-checking and translation bias, leading to a philosophical difference about whether AI bias can be meaningfully addressed. Professor Orero’s research-oriented optimism contrasted with the audience member’s practical pessimism about persistent bias, revealing different perspectives on the potential for technological solutions to social problems.


Topics

Human rights principles | Multilingualism | Gender rights online


Overall assessment

Summary

The session showed minimal disagreement among speakers, with most differences emerging from audience interactions rather than between the main presenters. The primary areas of disagreement centered on approaches to AI bias mitigation and strategies for AI education in universities.


Disagreement level

Low level of disagreement with significant implications for implementation strategies. While speakers agreed on fundamental goals of accessibility and inclusion, the differences in approach to persistent challenges like AI bias and educator training suggest that the field may need multiple complementary strategies rather than single solutions. The disagreements highlight the gap between research optimism and practical implementation challenges.


Partial agreements

Partial agreements

Similar viewpoints

Both speakers emphasize the importance of institutional control and awareness of systemic issues in AI systems – Orero focusing on bias recognition and Farrus on data privacy and security in educational contexts.

Speakers

– Pilar Orero
– Mireia Farrus

Arguments

The system is self-managed by universities to ensure privacy and security of data and content


The Alfie project addresses biases in media and AI models, particularly cultural and linguistic biases in Europe


Topics

Privacy and data protection | Human rights principles | Online education


Both speakers recognize that accessibility is not one-size-fits-all but requires nuanced approaches with different levels of simplification and support materials tailored to specific user needs and contexts.

Speakers

– Ana Matamala
– Estel la Oncins

Arguments

Easy-to-understand language exists on a continuum from easy language to plain language, addressing different user needs


Implementation requires consideration of time constraints, easy language needs, and provision of storyboard templates for storytelling activities


Topics

Rights of persons with disabilities | Online education | Digital access


Both speakers frame accessibility work as a matter of justice and empowerment, emphasizing participatory approaches and the role of academic research in addressing real-world social problems.

Speakers

– Mar Gutierrez Colon
– Anita Lamprecht

Arguments

Digital accessibility represents educational justice, requiring design with learners rather than for learners


Female academic leaders demonstrate how university research can solve real-world problems through AI applications


Topics

Human rights principles | Gender rights online | Capacity development


Takeaways

Key takeaways

AI can significantly enhance media accessibility through services like subtitling, audio description, and sign language interpretation, with projects like Mosaic creating centralized hubs for European broadcaster archives


Easy-to-understand news exists on a continuum from easy language to plain language, with 13 out of 39 surveyed European broadcasters already producing some form of simplified news content


Virtual reality shows strong potential in education, particularly for neurodiverse students, with both teachers and students responding positively to VR implementation in learning environments


Real-time transcription and translation systems like Scribal can improve accessibility in university settings, especially for minority languages and students with hearing impairments


Digital accessibility represents educational justice, requiring design approaches that work with learners rather than for them


Female academic leaders are successfully bridging university research with real-world societal problems through AI-enhanced accessibility solutions


Collaboration between academia and industry is essential for effective knowledge transfer and addressing accessibility challenges


Bias in AI systems, particularly in translation and media representation, is persistent and reflects inherent human biases that are difficult to eliminate


Resolutions and action items

The ENACT project will publish survey results and report on their website for broader access


Future development of a guidebook for teachers implementing VR technologies in educational settings


Creation of a VR platform with embedded co-created social studies content


Development of implementation manuals and scientific reports on VR for addressing neurodiversity in education


Scribal system to be presented at the Speech conference in Rotterdam for wider dissemination


Continued pilot testing and scaling of accessibility tools to other universities and countries


Publication of research results from the Kenyan reading comprehension app study


Unresolved issues

How to effectively train university educators in rapidly evolving AI and quantum technologies when industry experts are concentrated in few companies and reluctant to teach


How to develop standardized curricula and syllabi for emerging technologies that change faster than traditional academic approval processes


How to eliminate or significantly reduce persistent biases in AI translation systems, particularly gender-related translation errors


How to balance journalistic principles with accessibility requirements in news production


How to make accessibility services a strategic priority for broadcasters rather than an afterthought


How to scale successful accessibility solutions globally, particularly for underserved communities and minority languages


Suggested compromises

Universities can employ industry professionals as part-time teachers (6 hours only) to bridge the gap between rapidly evolving technology and academic instruction


Relying on internet resources and AI tools like ChatGPT for technology education while acknowledging their limitations and potential data currency issues


Accepting that some level of bias in AI systems may be inevitable since they reflect human biases, while continuing efforts to minimize harmful impacts


Focusing on collaboration between academia and industry at multiple levels rather than expecting complete solutions from either sector alone


Thought provoking comments

Digital accessibility is educational justice

Speaker

Dr. Mar Gutierrez Colon


Reason

This statement reframes accessibility from a technical challenge to a fundamental rights issue, elevating the entire discussion from ‘nice to have’ features to essential educational equity. It connects technology development with social justice principles.


Impact

This comment provided a powerful conceptual framework that tied together all the presented projects under a unified ethical imperative. The moderator specifically highlighted this as impressive, showing how it resonated and provided a memorable closing theme for the session.


We must design with, not for learners

Speaker

Dr. Mar Gutierrez Colon


Reason

This challenges the traditional top-down approach to educational technology design, advocating for participatory design that includes the target users as co-creators rather than passive recipients. It represents a fundamental shift in design philosophy.


Impact

This comment introduced a critical design principle that contrasts with many technology development approaches, emphasizing user agency and participation in the creation process rather than assumptions about user needs.


You will never eliminate bias, whatever you want… AI reflects who we are, reflects humans, and humans are biased

Speaker

Professor Pilar Orero


Reason

This is a brutally honest acknowledgment of AI’s fundamental limitations, challenging the often optimistic narratives around AI solutions. It introduces philosophical depth about the nature of technology as a mirror of human society.


Impact

This response to a question about bias in translation shifted the discussion from technical solutions to fundamental philosophical questions about AI’s relationship to human nature. It grounded the conversation in realistic expectations while maintaining the importance of the work.


How do we teach them in the university colleges because we don’t get the professors or trained people and technology developers of the world, they do not teach

Speaker

Professor NK Goel


Reason

This question exposed a critical gap between technological advancement and educational capacity, highlighting the practical challenges of implementing AI education at scale. It brought real-world constraints into the academic discussion.


Impact

This intervention shifted the focus from showcasing successful projects to addressing systemic challenges in technology education. It prompted responses about industry-academia collaboration and revealed the scalability challenges facing educational institutions globally.


Information should be for all, and news are a fundamental part of this information. And they are also a very fundamental part of participation

Speaker

Professor Ana Matamala


Reason

This connects media accessibility to democratic participation, framing accessible news not just as a service but as essential for civic engagement. It elevates the stakes of the technical work being presented.


Impact

This comment established the democratic and civic importance of the technical projects being discussed, providing a broader social context that connected individual accessibility needs to societal participation and inclusion.


Climate change presents data in a very unaccessible way… scientists are producing fantastic data, but the citizen is not getting that information

Speaker

Professor Pilar Orero


Reason

This insight reveals how accessibility challenges extend beyond traditional disability considerations to affect public understanding of critical issues like climate change. It demonstrates the broader societal implications of accessible design.


Impact

This comment expanded the scope of accessibility beyond the expected domains (education, media) to show how accessibility principles apply to urgent global challenges, demonstrating the universal relevance of their work.


Overall assessment

The key comments transformed this session from a series of project presentations into a deeper exploration of the philosophical, ethical, and practical dimensions of accessible technology. The discussion evolved from technical demonstrations to fundamental questions about justice, bias, and the relationship between technology and society. The audience questions, particularly about educational challenges in implementing AI, grounded the academic presentations in real-world constraints and sparked honest discussions about limitations and systemic barriers. The most impactful comments reframed accessibility as a justice issue and acknowledged the inherent challenges in AI development, creating a more nuanced and realistic understanding of both the potential and limitations of technology in promoting inclusion.


Follow-up questions

How can universities effectively integrate industry experts from emerging technologies like AI and quantum computing into their teaching programs when these experts are limited and concentrated in only a few major companies?

Speaker

Professor NK Goel from India


Explanation

This addresses a critical gap in higher education where traditional industry-academia collaboration models don’t work for cutting-edge technologies due to the scarcity and exclusivity of experts in major tech companies.


What are the specific types and extent of bias present in AI translation systems, particularly regarding gender and other subtle linguistic biases, and how can these be mitigated?

Speaker

Tamiris (audience member working with fact-checking)


Explanation

This is crucial for ensuring accuracy and fairness in AI-powered translation and communication tools, especially given the persistent nature of gender bias in translation systems.


How can educators be systematically upskilled to teach new technologies when there are no established courses, syllabi, or approval systems for emerging tech subjects?

Speaker

Professor NK Goel from India


Explanation

This highlights the structural challenges in educational institutions that prevent rapid adaptation to technological changes and the need for new frameworks for curriculum development in emerging fields.


What are the detailed survey results and percentages from the ENACT project’s mapping of easy-to-understand news practices across European broadcasters?

Speaker

Ana Matamala


Explanation

The presenter mentioned they would publish detailed survey results on their website, indicating there’s more comprehensive data available that could inform best practices in accessible news production.


How can the interaction between accessibility requirements and journalistic principles be effectively balanced in news production?

Speaker

Ana Matamala


Explanation

This was identified as a key challenge in the ENACT project, requiring further exploration to develop guidelines that maintain journalistic integrity while ensuring accessibility.


What specific training methodologies and resources are most effective for helping educators integrate VR technologies into curricula for neurodiverse students?

Speaker

Estella Oncins


Explanation

The research identified a clear need for comprehensive training tools, but the specific approaches and their effectiveness require further investigation.


How can time allocation and session duration be optimized when implementing VR co-creation activities with students in educational settings?

Speaker

Estella Oncins


Explanation

The pilot workshop revealed that standard time allocations (1.5 hours) were insufficient for VR activities, indicating need for research on optimal scheduling for immersive learning experiences.


Disclaimer: This is not an official session record. DiploAI generates these resources from audiovisual recordings, and they are presented as-is, including potential errors. Due to logistical challenges, such as discrepancies in audio/video or transcripts, names may be misspelled. We strive for accuracy to the best of our ability.