Lightning Talk #90 Tower of Babel Chaos

Summary

This discussion focused on exploring communication barriers in internet governance by experimenting with multilingual participation rather than defaulting to English as the common language. Virginia Paque, the session moderator, initiated the experiment by suspending the rule of English as the universal language, despite being the only native English speaker among nearly two dozen participants. The session began with various speakers defining internet governance in their preferred languages, with most referencing the WGIG definition that emphasizes multi-stakeholder collaboration between governments, private sector, and civil society.

The core experiment involved participants communicating in their native languages to simulate a “Tower of Babel” scenario and observe what challenges and solutions emerged. Several participants reported feeling confused and migraine-inducing chaos when unable to understand others, with people naturally clustering into language groups with speakers they could comprehend. Some participants discovered they were the only representatives of their languages present, including speakers of Maltese, Samoan, Cape Verdean Creole, and Chichewa.

Ken Huang from Lingo AI presented artificial intelligence as a potential solution, explaining that AI can theoretically process all 7,000 human languages but defaults to English and major languages when data sets are insufficient. Other technological solutions discussed included real-time translation devices and the need for better multilingual datasets. Participants debated whether English should remain the de facto international language due to its practical effectiveness, or whether multiple language options should be provided to increase inclusivity.

The discussion revealed the political and cultural complexities of language choice, with examples from India and China where English serves as a neutral option among competing local languages. The experiment ultimately highlighted both the necessity of common communication methods and the potential for technological solutions to bridge linguistic divides in global internet governance discussions.

Keypoints

**Major Discussion Points:**

– **Language barriers in global internet governance discussions** – The session explored how requiring English as a common language excludes voices and creates communication challenges, despite most participants being non-native English speakers

– **Experimental multilingual communication approach** – Participants engaged in a “Tower of Babel” experiment where English was suspended as the required language, allowing people to communicate in their native languages to observe what happens

– **Technology solutions for language barriers** – Discussion of AI translation capabilities, with insights that AI thinks in mathematics rather than any specific language, and exploration of real-time translation tools and their current limitations

– **The politics and practicality of language choice** – Debate over whether English should remain the default international language due to its practical effectiveness versus concerns about linguistic imperialism and the need for more inclusive multilingual options

– **Isolation of minority language speakers** – Recognition that many participants were the sole representatives of their native languages (Maltese, Samoan, various Creoles, etc.), highlighting the challenge of meaningful participation in global forums

**Overall Purpose:**

The discussion aimed to examine and challenge the dominance of English in international internet governance forums by experimenting with multilingual communication approaches and exploring technological and policy solutions to make global discussions more linguistically inclusive.

**Overall Tone:**

The tone began as experimental and somewhat chaotic during the multilingual exercise, with participants reporting confusion and difficulty communicating. However, it evolved into a thoughtful, collaborative discussion as participants shared insights from the experiment. The atmosphere remained respectful and constructive throughout, with genuine curiosity about finding solutions to language barriers, though some participants ultimately concluded that English remains practically necessary for international communication.

Speakers

**Speakers from the provided list:**

– **Virginia (Ginger) Paque**: Representative of CADE consortium, native English speaker, session moderator and organizer

– **Abed Kataya**: Internet governance expert (no specific title mentioned)

– **Kenneth Harry Msiska**: Internet governance expert, references WGIG (Working Group on Internet Governance)

– **Stephanie Borg Psaila**: Maltese speaker, appears to be involved in international governance forums

– **Karolina Iwańska**: Internet governance expert (no specific title mentioned)

– **Slavica Karajicic**: Internet governance expert (no specific title mentioned)

– **Bimsara Malshan**: Internet governance expert (no specific title mentioned)

– **Ken Huang**: Co-founder of Lingo AI, from Singapore Internet Governance Forum

**Additional speakers:**

– **Audience members** (multiple unnamed participants who shared observations during the discussion, including speakers of various languages such as Chinese, German, Samoan, Hindi, Cape Verdean Creole, Chichewa, and others)

– **Una**: Participant from China, involved in language technology research and community language projects

# Report: Experimental Discussion on Multilingual Communication Barriers in Internet Governance

## Executive Summary

This experimental session, moderated by Virginia Paque, explored communication barriers in internet governance forums through a unique “Tower of Babel” approach. The session brought together participants from diverse linguistic backgrounds to examine challenges and potential solutions for multilingual participation in global digital policy discussions. The experiment involved temporarily suspending English as the universal language to allow participants to experience firsthand the communication barriers that typically remain hidden when English dominance is accepted as standard practice.

The session revealed tensions between linguistic inclusivity and practical communication needs, highlighting the complex relationship between language, technology, and global governance. Participants discussed questions of fairness, efficiency, and the role of emerging technologies in bridging linguistic divides.

## Participant Overview and Experimental Context

The discussion featured participants from various linguistic backgrounds, with Virginia Paque serving as moderator. As she noted, “I have spent most of my life speaking Spanish although English is my native language.” Key participants included Abed Kataya, Kenneth Harry Msiska, Stephanie Borg Psaila from Malta, Karolina Iwańska, Slavica Karajicic, Bimsara Malshan, and Ken Huang, co-founder of Lingo AI from the Singapore Internet Governance Forum.

The session began with what appeared to be a reference to “COVID-19 Refugee Camp Assisted Living” before transitioning to the main discussion. Participants represented languages including Chinese, German, Samoan, Hindi, Cape Verdean Creole, Chichewa, Swahili, and others, creating a genuinely multilingual environment for the experiment.

## Definitions of Internet Governance

Before the multilingual experiment, participants provided definitions of internet governance in their preferred languages. Abed Kataya emphasized comprehensive collaboration, defining it as involving cooperation between government, private sector, civil society, and technical communities. Kenneth Harry Msiska referenced the WGIG definition, describing it as establishing rules, policies, and procedures applied jointly by all stakeholders.

Stephanie Borg Psaila offered a different perspective, critiquing the terminology itself for overemphasizing government roles with insufficient attention to civil society participation. Other participants provided complementary definitions: Karolina Iwańska emphasized decentralized management, Slavica Karajicic highlighted the multidisciplinary nature encompassing infrastructure, standards, security, law, economics, development, culture, and human rights, and Bimsara Malshan focused on shared principles and decision-making procedures.

## The Tower of Babel Experiment

The session’s central experiment involved encouraging participants to communicate in their native languages to observe emerging challenges and solutions. Virginia Paque initiated this experiment despite being a native English speaker, stating her goal was to highlight issues with English dominance.

The immediate results varied among participants. Some reported confusion during the multilingual phase, with one audience member describing the experience as “chaotic” and noting they could only connect with Swahili speakers. However, Virginia Paque herself observed that it was “not that chaotic.”

Participants naturally began clustering into linguistic groups, seeking speakers of languages they could understand. Several participants discovered they were the sole representatives of their native languages at the forum, including Stephanie Borg Psaila as the only Maltese speaker and others representing Samoan, Cape Verdean Creole, and Chichewa.

## Technological Solutions and AI Capabilities

Ken Huang presented insights on artificial intelligence capabilities, explaining that AI can theoretically process all 7,000 human languages but defaults to English and major languages when data sets prove insufficient. He noted that “AI can design their own Native computing languages” distinct from human languages, suggesting possibilities for communication systems that transcend traditional linguistic boundaries.

An audience member added that AI “thinks in mathematics and digit proximity” rather than any specific language, making it potentially culturally neutral. However, participants also noted limitations, with current AI speech recognition handling only about 100 languages with limited effectiveness, and 95% of internet language data existing in English.

The discussion revealed that Google Translate is adding “100 languages every year,” showing progress in technological solutions while acknowledging current constraints.

## The English Dominance Discussion

The session revealed different perspectives on English as a common language. After experiencing the multilingual experiment, one participant concluded that “English is the solution for the chaotic Tower of Babel situation.” A Hindi-speaking participant explained that in India’s multilingual context, English serves as a politically neutral option when native languages carry political implications.

Abed Kataya provided historical context, noting that English is “the third most spoken native language globally” following Chinese Mandarin and Spanish, but serves as the current business language due to power structures. He suggested that language dominance follows historical patterns, with different languages serving as lingua francas in different eras.

However, Virginia Paque questioned the fairness of requiring English when most participants are non-native speakers. Stephanie Borg Psaila proposed alternative approaches, suggesting multiple language options with simultaneous interpretation, similar to UN and EU practices.

## Cross-Linguistic Communication Approaches

The discussion explored alternatives to the binary choice between English dominance and multilingual chaos. Participants identified cross-linguistic communication as a promising approach, where speakers of related languages can communicate in their native tongues while understanding responses in different but related languages. Examples mentioned included “Portunol, Spanglish” as forms of cross-linguistic communication.

One observation noted that when Spanish and Hindi speakers attempted to communicate, some words were “close enough to English” to facilitate understanding, demonstrating natural bridges between languages.

Virginia Paque also raised the topic of internationalized domain names as another aspect of multilingual internet governance that requires consideration.

## Areas of Agreement and Disagreement

Participants demonstrated consensus on several issues, including the need for multi-stakeholder collaboration in internet governance and acknowledgment that language barriers create challenges in international forums. There was also general agreement that AI offers potential for addressing language barriers while facing current limitations.

However, significant disagreements emerged regarding solutions. The most notable disagreement concerned whether to maintain English as the universal solution or implement multiple language options. Stephanie Borg Psaila advocated for multiple language choices with interpretation, while others defended English as practical and effective.

Disagreements also emerged about AI capabilities, with varying levels of optimism about technological solutions versus emphasis on current limitations.

## Key Insights and Observations

The experimental approach provided several insights into multilingual communication challenges. The session demonstrated that language barriers create genuine difficulties in international forums, though the severity of these challenges varied among participants. The experiment showed how participants naturally seek linguistic connections and form communication clusters.

The discussion highlighted both the practical effectiveness of English as a common language and concerns about fairness when most participants are non-native speakers. Technological solutions emerged as promising but currently limited, particularly for less common languages and oral communication.

The session also revealed that different participants have varying tolerance for multilingual communication challenges, with some finding creative ways to bridge language gaps while others prefer clear common language solutions.

## Conclusion

This experimental session provided insights into the complex challenges of multilingual communication in internet governance forums. By temporarily suspending English dominance, participants experienced linguistic barriers firsthand and explored various approaches to multilingual communication.

While no definitive solutions emerged, the discussion revealed the trade-offs between inclusion and practicality, the potential and limitations of technological solutions, and the varying perspectives on language choice in international forums. The session demonstrated that addressing communication barriers requires balancing practical communication needs with concerns about fairness and inclusion.

The experiment highlighted that meaningful progress on linguistic inclusion may require willingness to experiment with established practices while acknowledging both practical constraints and equity concerns in international governance processes.

Virginia (Ginger) Paque: COVID-19 Refugee Camp Assisted Living Good morning, buenos dias. Hola, como estan? Soy Ginger para el consorcio CADE, where we will all be working together. We’re starting with the proposition that the largest, strongest challenge to multi-stakeholder inclusion and voices in global processes is communication. This challenge predates the digital divide. It underlies the digital divide. So we will now try to work with that problem because principle two resolving that challenge of communication is the fact that the biggest challenge to communication is language. We have been communicating in CAID in English in spite of the fact that out of almost two dozen people, I am the only person who is a native English speaker. Is that fair? We are proposing now for this session to start with a basic discussion of internet governance without the stipulation of English as the imposed common language. So the rule of English as a common language is now suspended. I can speak English because it’s my native language. If English is internative language, there are no rules. There’s a suspended rule. So I invite you all to participate. We will start with our definitions and I invite my colleagues who will start the discussion with internet governance in their own languages and then we will open the floor and ask each of you to be a participant, a panelist, and an active member. Thank you very much and welcome.

Abed Kataya: The internet governance is the development and implementation of the comprehensive government and private sector, civil society and technical society, all in turn, of the principles, standards, rules, and procedures for decision-making and joint activities that form the development of the internet and its use and the setting of programs that define the form of the internet and its use. The internet governance is an essential issue because of the internet’s potential to enhance the sustainable development of humanity and the building of comprehensive knowledge societies and to enhance the free flow of information and ideas all over the world. And now I leave you with my colleague, Kenneth Harry Msiska.

Kenneth Harry Msiska: Thank you very much. According to the WGIG (Working Group on Internet Governance), governance means establishing rules, policies, and appropriate procedures that are applied jointly by all stakeholders—such as governments, companies, and non-governmental organizations—while respecting principles, frameworks, laws, and decision-making processes, as well as policies that promote governance and its effective implementation.

Virginia (Ginger) Paque: The WIEGIG definition of internet governance stipulates that internet governance is the development and application by governments, the private sector, and civil society in their respective roles of shared principles, norms, rules, decision-making procedures, and programs that shape the evolution and the use of the internet. The WIEGIG definition of internet governance stipulates that internet governance is the development and application by governments, the private sector, and civil society in their respective roles of shared principles, norms, rules, decision-making procedures, and programs that shape the evolution and the use of the internet.

Stephanie Borg Psaila: Internet governance focuses on the role of government in shaping digital policy. The phrase places importance on the role of government in governance, but less on the role of civil society. development and application by governments, the private sector, and civil society in their

Karolina Iwańska: Internet governance uses the same word we would use to describe managing a company, a team, or a crisis. So “Internet governance” emphasizes the decentralized nature of the Internet, rather than its focus on regulations or government institutions.

Slavica Karajicic: Internet governance is the development and application of common principles, norms, rules, decision-making processes and programs that shape the evolution and use of the Internet, by governments, the private sector and civil society, in their specific roles. It is a multidisciplinary field, encompassing many aspects: infrastructure, standards, security, law, economics, development, culture, human rights, etc. (Dictionary of Internet and Communications)

Bimsara Malshan: respective roles of shared principles, norms, rules, decisions-making procedures, and programs that shape the evolution and the use of the internet. The WIEGIG definition of internet governance stipulates that internet governance is the development and application by governments, the private sector, and civil society in their respective roles of shared principles, norms, rules, decisions-making procedures, and programs that shape the evolution and the use of the internet.

Virginia (Ginger) Paque: Thank you very much for listening, we now invite everyone in yours place or get up and join us, the closer you are to the stage, the better talk, the more the camera will capture this experiment in vivo. Again, I’m speaking English because it is my native language, I invite you to remember that we have we suspended English as a common or universal language i invite you to use whatever language, gesture or way of communication that works, we are trying to see what hepen to see what whoul hepeneds if we had honest tower of babel. Would you please join us, I know we have other languages… To simulate what is, how are we going to join silos… There are some questions we are going to consider. This right now is a group exercise, we don’t have individual speakers, we will have individual speakers after this group exercise, so this is a group exercise, so let’s experiment. Okay, I’m hearing a lot of things going on Because of the chaos, which is not that chaotic But I’ve heard a lot of people not understanding each other I invite you to take your places or to stay standing We’ll do a little bit more directed analysis now Where we have people speaking in their own languages Or English or however you want To communicate the best you can As we work out, for instance, any of the questions If I can ask Slavica to please put the questions back on the screen Because that’s where we’ll be going now I hope you have the basic context from our cheat sheet I invite, we have four microphones So if people want to speak, come around But come up this way so the camera can catch you I invite you to tell us your suggestions For instance, I had a gentleman who was I was very pleased, was representing the Chinese language Where are you? He was proposing actually an AI solution I hope you will come up and join us at the mic I invite you to all come up with 30-second interventions Especially representing languages that have not been used And if you have a solution, what worked right now? What didn’t work? Anybody talking sign language? Did anyone find themselves saying, i can’t understand what you’re telling me? So what happened? Come and tell us. 30 seconds each. Come on up. Come and tell us about your solution. You have 30 seconds. Anybody who wants to, mic is open, we have four mics 30 Seconds each to tell us your answers Or your comments. Please go ahead.

Ken Huang: I’m from lingo AI I’m the co-founder We are from Singapore Internet Governance Forum We are also the co-founders So what language does AI think in? AI is thinking AI can think in every language All 7,000 languages But if we don’t have enough data sets Then it thinks in English And other major languages

Virginia (Ginger) Paque: How interesting So we have a default language for AI as well

Ken Huang: You know, AI actually can talk in Other than the 7,000 languages They can design their own Native computing languages It’s different from human languages

Virginia (Ginger) Paque: And that is such an interesting concept Which is why i invite you all to visit the booth And find out more about this project Because is AI our solution? Is that how we’re going to communicate in the future? What does AI do? So please join the booth to find out more about that possibility Because it’s one of our possible solutions What else do you think? What did you find out? What’s your observation from that flash experiment?

Audience: Excellent experience. Thank you for organizing it It was easy and interesting to find a conversational partner Due to the label So labeling the label of the language that people speak Because majority of people here speak more than one language And we could have had a list of languages On our chests To easily find more communication languages And surprisingly you can find people that speak

Virginia (Ginger) Paque: What would be the common language? That is very interesting We did encourage people to put their native language Their first language to see the diversity of primary languages But that is an excellent suggestion Because we would have helped direct those silos First of all Which would have formed if we find commonalities And we might also have found That there is an alternative second language Or common language So yes, I like that. Next time we’ll have to try it This is our first experiment We need to know what did you think? What did you hear? What did you find out?

Audience: I have an answer here From the experiment I almost got a migraine Everyone was speaking whatever I was not understanding I was only picking from the Swahili speakers So those are the African languages And from Uganda I think we have the Bantu Or the two languages so they sort of overlap The Spaniards and the rest It’s chaotic

Virginia (Ginger) Paque: Does anyone dare to share?

Audience: Thank you for the experiment

Virginia (Ginger) Paque: Can you come up so the camera might catch you? Walk up while you’re talking Keep going

Audience: So I saw that during the experiment people from the same nation would gather up Together in small groups like the English speakers I was speaking Chinese with my Chinese fellows I feel like that separates us as a nation Because we would only be in groups of the people we understand So I do think English as a common language Does work in an international context English is the solution for the chaotic Tower of Babel situation So I do think English is the solution for the chaos

Virginia (Ginger) Paque: Thank you What is your native language?

Audience: German but I do speak Chinese

Virginia (Ginger) Paque: So you chose and Chinese is one of the candidates for a common language Except that I have given conferences in situations for Chinese people Who have the common language of English Because they don’t understand each other That happens also in Africa and parts of the US So Samoa, would you please join us? Thank you very much

Audience: Thank you To me it was a confusing one Because I’m the only Samoan here And only me understand my language So in this global platform It should be a common language For all of us can understand each other So as we know that English we say is an international language So we need to learn that so that we can communicate There are some exceptions It depends on the target and the market And the number of people in that space And then you can have that common language Or your native language to communicate But for international ones like here It should be a common language So that we can easily communicate

Virginia (Ginger) Paque: Do I have anyone waiting?

Audience: Hi, so I was talking to the gentleman I don’t know if he’s here He was talking in Spanish and I was talking in Hindi But the thing was some of the words that he was saying Were close enough to English that I could kind of understand So that’s there And then the second thing is in Hindi I don’t even know I looked up what the word for internet is I don’t think there is a word So some of these newer terms there are not even native terms for them that exist And then just coming back to the point of India has so many dialects We have no national language because language is very political And there are some languages like Hindi That seem to be the national language Then there’s pushback from other regions because of culture So English becomes the default language Because it’s politically the most equal So that’s just one point

Virginia (Ginger) Paque: So I have a couple of things Please pardon me for monopolizing the mic And raise your hand when you want to join in Come on up in the meantime English is the de facto second language And perhaps is the best solution because it works There are those who would say English is an imposed language And you use the word political So I ask you to consider And add your comments If you think English has been imposed Or yes, it’s imposed, but it works The other thing I would like to also ask Is how many of you Think you’re the only speaker of your language We know of two, we think Only one Maltese speaker, only one Samoan Only one Polish What language is yours? Can you get to a mic? Can you come up?

Audience: I speak two types of Creole And I think I’m one of the few Cape Verdeans here In Oslo

Virginia (Ginger) Paque: Excellent, so we now have at least four people Who represent, they’re the only representative Of their language That’s a challenge that they have obviously overcome And please feel free to comment And please look for a mic or come up to speak We want to hear your reactions

Audience: I’m from Malawi, and my local language is Chichewa But according to the experience What I try to do is just to use Geese as one of the Speakers of my language And I think that’s a challenge And I think that’s a challenge And I think that’s a challenge And I think that’s a challenge as well as the sign language because I could see like there are different tags so when I look at the person it was just pointing at the language so I could understand like he was telling me like I’m coming from this particular region and what’s your name where are you coming from so it was all about guessing and also about sign language but the experience was so fun and it taught me something that language is the best thing to do more especially in international gatherings like this we need to have a common language common ground which we can use but just like as the portrait of the Tower of Babel if we have different languages it is so hard for us to understand each other and also to pursue a common goal thank you

Virginia (Ginger) Paque: how many people um anyone ready is anyone waiting for a mic please take the mic this is you you are the panelists where are my panelists

Stephanie Borg Psaila: can I hi I’m not only the only one who speaks Maltese I’m probably the only Maltese here right I have never been to an IGF with maybe one or two exceptions where there were other people from Malta at the IGF wherever it’s been organized so I’m a bit isolated but English breaks down that isolation strangely enough it was in Italian that we communicated so we found a second language there or a third I would say I want to challenge the notion of one solution of English being one solution and I would add to our colleague from the Council of Europe who was mentioning choices and I think this is actually the way forward of having choices it is why the UN has several official languages it is why the EU has multiple official language so perhaps rather than saying let’s have English as a common language why not give people more options I’m not saying 1000 option to choose from because that would be taking it a bit too far right but what if in the discussions that we have there are more simultaneous interpretations in at least a handful of different languages that people have choices which language to follow I think we would tick more boxes in terms of participation of people if there were even more languages I’m not expecting any fora to choose Maltese as another possibility that would be wouldn’t make a lot of sense but there are so many other languages which are you know a lot of people know Spanish, Chinese, Swahili, German, French, so many

Virginia (Ginger) Paque: and your point of those of us we can talk to Italians I should explain I spent most of my life speaking Spanish although English is my native language and we have Portunol, Spanglish, well there are many more combinations where we actually do speak our own languages and we are answered in the other language which we can understand so Stephanie that’s an exciting possibility of looking for channels joining in together would you have a comment on that you have you deal with a lot of languages and you had an interesting

Audience: recently I was in a also interesting presentation on large language models and I learned that AI doesn’t think in any language it thinks in terms of maths mathematics the proximity of different notions which are coded in digits so indeed let’s not how to say be fooled by the idea that one language is going to predominate it is mathematics actually all digital or everything which is digital are digits zero or one I think that’s it so it is important to use this platform as I mean the digital world and the AI as a neutral culturally relatively neutral of course it feeds on existing data but that evolves quickly and basically I would say let’s trust and explore this world.

Virginia (Ginger) Paque: and hope that those who are writing the algorithms speak a lot of languages what one I’m sorry for this bad joke I’m going to make at least in digital and binary we only have to learn two letters it’s not like these languages my colleagues just have been speaking that I can’t even read zero and one I can understand anyone else on that and I would go back to a comment do you have a response to that my friend that the AI thank you very much for that addition I think it’s very important to our look search for solutions

Audience: hello everyone my name is Una and I came from China you know in China there’s the official language is Mandarin but actually there are around 60 local languages and nuances and they speak in different ways of the communication so once the Chinese people learning English and when they go to the international platform they are not going to speak and communicate very well because there’s no such very like more practice every day in English so we see that the problem is that not languages is like a barriers it’s not like a barrier not languages is like a barriers it’s not like bridges they can connect everyone that we are in the IGF and we speak different languages around hundreds of languages that hundreds of countries that we are from but we can only speak in English to communicate so we are now do some consumption it’s like how we can connect like speak with our own native languages and we use like some hardwares like earphones can translate all the languages into other native languages let’s see I speak Chinese you speak English and we can understand well in each other native languages so we know that Google Translate has added 100 languages every year for inclusive language development but actually there’s some text to text translation but not on the oral communication so kind of the technologies like AISR can only recognize 100 languages but not in a very high quality that means you cannot rely on that technology so we lack of the data sets because almost 95% of data sets or languages data in the internet is English so the rest of languages and the people are not communicating in their own internet language in the internet that’s the problem so the current logical models understand in English or Spanish or Chinese but not all the other languages so there leads to the buyers that means AI cannot understand your own culture your own native languages or your own native like local like the knowledge there so language is very complex yeah so we are now doing some project research on like community on the community language community side that means all the people from all the world they can contribute their native knowledge in their own native language.

Virginia (Ginger) Paque: that’s amazing that that’s a very eye-opening insight and we really appreciate that especially added to others on that note you mentioned technology and both the advantages of and shortcomings which was excellent I did wonder if we would see people pulling out their phones for Google Translate to help them during the discussion I did not see any cases of that so that’s very interesting if there are techies in the group I I’m always curious whether the use of the internationalized domain names might have so because they they have to rationalize all of these different languages and it comes into I think a combination rather than a dominance because we do of course need our our domain names in languages we can comprehend or we can’t use the internet I have someone else

Abed Kataya: oh yeah it’s me oh so actually uh let me uh explain something that English is the third most spoken language native spoken language in the world it’s not the first one uh the first one is the Chinese Mandarin the second one is Spanish the third one is the English and I think why we are speaking English now especially in the business it’s a business language because it is uh yeah because it is uh like the dominant uh the dominant powers language that’s why and every era has its own language I mean let’s imagine that Arabic has used to be like the dominant language before then Spanish in some eras and then that like so I think maybe next we’re gonna speak Chinese Mandarin we don’t know as our business language.

Virginia (Ginger) Paque: well we definitely have a way forward I would love anyone who can to take but we do need to close the the timer’s yelling at me in red um which makes sense I thank you all for your input we apologize for the chaos but the chaos was important certainly among our thank yous thank you to all of you and we should have a slide with thank you in many languages but I don’t see it thank you to the tech team for supporting us thank you for the online participants who are watching us and can’t raise their voices in whatever language they want thank you all very very much an applause for you

Agreement points

Internet governance requires multi-stakeholder collaboration

Speakers

– Abed Kataya
– Kenneth Harry Msiska
– Slavica Karajicic
– Bimsara Malshan

Arguments

Internet governance involves comprehensive collaboration between government, private sector, civil society and technical communities in developing principles and standards

Internet governance establishes rules, policies and procedures applied jointly by all stakeholders while respecting frameworks and decision-making processes

Internet governance is multidisciplinary, encompassing infrastructure, standards, security, law, economics, development, culture and human rights

Internet governance involves shared principles, norms, rules and decision-making procedures that shape internet evolution and use

Summary

All speakers agree that internet governance fundamentally requires collaboration among multiple stakeholders including governments, private sector, civil society, and technical communities to develop shared principles, norms, and decision-making procedures

Topics

Legal and regulatory

Language barriers create significant challenges in international forums

Speakers

– Virginia (Ginger) Paque
– Stephanie Borg Psaila
– Audience

Arguments

Communication challenge predates and underlies the digital divide, with language being the biggest barrier to communication

Some participants are the only speakers of their native language at international forums, creating isolation

Chaos and confusion result when people cannot understand each other in multilingual settings

Language separation creates national silos where people only communicate within their linguistic groups

Summary

Speakers consistently acknowledge that language barriers pose fundamental challenges to effective communication and participation in international internet governance discussions

Topics

Sociocultural

Similar viewpoints

Both speakers challenge the dominance of English as the sole common language and advocate for more inclusive multilingual approaches in international forums

Speakers

– Virginia (Ginger) Paque
– Stephanie Borg Psaila

Arguments

English dominance is unfair when only one participant is a native English speaker among dozens

Multiple language options should be provided rather than imposing one common language, similar to UN and EU practices with official languages

Topics

Sociocultural

Technology and AI solutions have potential for addressing language barriers but currently face significant limitations due to data bias toward English and major languages

Speakers

– Ken Huang
– Audience

Arguments

AI can theoretically think in all 7,000 languages but defaults to English and major languages when data sets are insufficient

AI thinks in mathematics and digit proximity rather than any specific language, making it culturally neutral

Current translation technology lacks high-quality oral communication capabilities and sufficient data sets for most languages, with 95% of internet language data being in English

Topics

Infrastructure

English functions as a practical common language solution despite not being the most widely spoken native language, serving as a politically neutral option in complex multilingual contexts

Speakers

– Abed Kataya
– Audience

Arguments

English is the third most spoken native language globally, following Chinese Mandarin and Spanish, and serves as the current business language due to dominant power structures

English serves as the de facto second language and works effectively as a practical solution

English functions as a politically neutral default language in multilingual contexts like India where native languages are politically charged

Topics

Sociocultural

Unexpected consensus

English as both problem and solution

Speakers

– Virginia (Ginger) Paque
– Stephanie Borg Psaila
– Audience

Arguments

English dominance is unfair when only one participant is a native English speaker among dozens

English serves as the de facto second language and works effectively as a practical solution

English functions as a politically neutral default language in multilingual contexts like India where native languages are politically charged

English is necessary as a common language for international gatherings to pursue common goals

Explanation

Despite initial criticism of English dominance, there emerged unexpected consensus that English, while problematic, serves as an effective practical solution for international communication. Even those who challenged its dominance acknowledged its utility

Topics

Sociocultural

Technology limitations despite AI potential

Speakers

– Ken Huang
– Audience

Arguments

AI can theoretically think in all 7,000 languages but defaults to English and major languages when data sets are insufficient

Current translation technology lacks high-quality oral communication capabilities and sufficient data sets for most languages, with 95% of internet language data being in English

Explanation

Despite presenting AI as a potential solution, there was unexpected consensus that current technology actually reinforces language inequalities due to data bias, making it less viable as an immediate solution than initially suggested

Topics

Infrastructure

Overall assessment

Summary

The discussion revealed strong consensus on the fundamental challenges of language barriers in internet governance and the need for multi-stakeholder collaboration, but also unexpected agreement that English, despite its problematic dominance, remains the most practical current solution

Consensus level

High consensus on problem identification and moderate consensus on solutions. The implications suggest that while participants recognize the inequity of English dominance, they also acknowledge practical constraints that make immediate alternatives difficult to implement. This creates a tension between idealistic multilingual goals and pragmatic communication needs in international internet governance forums.

Different viewpoints

English as the universal solution vs. multiple language options

Speakers

– Stephanie Borg Psaila
– Audience

Arguments

Multiple language options should be provided rather than imposing one common language, similar to UN and EU practices with official languages

English serves as the de facto second language and works effectively as a practical solution

English is necessary as a common language for international gatherings to pursue common goals

Summary

Borg Psaila advocates for multiple language choices with simultaneous interpretation in several languages, while audience members argue that English works effectively as a single common language solution for international forums

Topics

Sociocultural

Fairness vs. practicality of English dominance

Speakers

– Virginia (Ginger) Paque
– Audience

Arguments

English dominance is unfair when only one participant is a native English speaker among dozens

English serves as the de facto second language and works effectively as a practical solution

English functions as a politically neutral default language in multilingual contexts like India where native languages are politically charged

Summary

Paque questions the fairness of English dominance when most participants are non-native speakers, while audience members defend English as a practical and politically neutral solution that works effectively

Topics

Sociocultural

AI language capabilities and limitations

Speakers

– Ken Huang
– Audience

Arguments

AI can theoretically think in all 7,000 languages but defaults to English and major languages when data sets are insufficient

Current translation technology lacks high-quality oral communication capabilities and sufficient data sets for most languages, with 95% of internet language data being in English

Summary

Huang presents AI as having broad language capabilities across 7,000 languages, while an audience member emphasizes significant limitations in current AI translation technology, particularly for oral communication and less common languages

Topics

Infrastructure

Unexpected differences

Definition emphasis in internet governance

Speakers

– Stephanie Borg Psaila
– Other speakers providing definitions

Arguments

Internet governance focuses primarily on government’s role in shaping digital policy with less emphasis on civil society

Internet governance involves comprehensive collaboration between government, private sector, civil society and technical communities in developing principles and standards

Internet governance emphasizes decentralized management rather than focusing on regulations or government institutions

Explanation

While most speakers provided standard multi-stakeholder definitions of internet governance, Borg Psaila uniquely critiqued the terminology itself for overemphasizing government roles. This was unexpected as it challenged the fundamental framing rather than just the content of internet governance definitions

Topics

Legal and regulatory

Historical context of language dominance

Speakers

– Abed Kataya
– Audience

Arguments

English is the third most spoken native language globally, following Chinese Mandarin and Spanish, and serves as the current business language due to dominant power structures

English serves as the de facto second language and works effectively as a practical solution

Explanation

Kataya’s historical perspective on language dominance cycles (Arabic, Spanish, English, potentially Chinese) was unexpected as it reframed the English dominance debate from a practical communication issue to a broader discussion of power structures and historical patterns

Topics

Sociocultural

Overall assessment

Summary

The main areas of disagreement center on language solutions for international communication, with fundamental tensions between fairness/inclusion versus practicality/efficiency, and between single-language versus multi-language approaches

Disagreement level

Moderate to high disagreement with significant implications. The disagreements reveal deeper tensions about power structures, cultural representation, and practical governance in international forums. These disagreements could impact policy decisions about language accommodation, technology investment priorities, and the fundamental approach to inclusive participation in internet governance processes

Partial agreements

Similar viewpoints

Both speakers challenge the dominance of English as the sole common language and advocate for more inclusive multilingual approaches in international forums

Speakers

– Virginia (Ginger) Paque
– Stephanie Borg Psaila

Arguments

English dominance is unfair when only one participant is a native English speaker among dozens

Multiple language options should be provided rather than imposing one common language, similar to UN and EU practices with official languages

Topics

Sociocultural

Technology and AI solutions have potential for addressing language barriers but currently face significant limitations due to data bias toward English and major languages

Speakers

– Ken Huang
– Audience

Arguments

AI can theoretically think in all 7,000 languages but defaults to English and major languages when data sets are insufficient

AI thinks in mathematics and digit proximity rather than any specific language, making it culturally neutral

Current translation technology lacks high-quality oral communication capabilities and sufficient data sets for most languages, with 95% of internet language data being in English

Topics

Infrastructure

English functions as a practical common language solution despite not being the most widely spoken native language, serving as a politically neutral option in complex multilingual contexts

Speakers

– Abed Kataya
– Audience

Arguments

English is the third most spoken native language globally, following Chinese Mandarin and Spanish, and serves as the current business language due to dominant power structures

English serves as the de facto second language and works effectively as a practical solution

English functions as a politically neutral default language in multilingual contexts like India where native languages are politically charged

Topics

Sociocultural

Key takeaways

Language barriers are a fundamental challenge in global internet governance that predates and underlies the digital divide

English dominance in international forums creates unfairness when most participants are non-native speakers

Multilingual communication without a common language leads to chaos and formation of linguistic silos where people only communicate within their language groups

Many participants at international forums are isolated as the sole representatives of their native languages

Internet governance is defined as a multistakeholder process involving governments, private sector, civil society, and technical communities in developing shared principles and standards

AI and technology offer potential solutions but currently default to English due to insufficient data sets in other languages, with 95% of internet language data being in English

English serves as a practical common language solution despite being imposed, functioning as a politically neutral option in multilingual contexts

Cross-linguistic communication is possible between related languages where speakers can understand each other while using their native tongues

Resolutions and action items

Participants were invited to visit the AI booth to learn more about technological solutions for language barriers

Suggestion to implement language labeling systems in future experiments to help people find common communication languages

Proposal to provide multiple language options with simultaneous interpretation in several major languages rather than imposing one common language

Community-based projects should be developed to allow people to contribute native knowledge in their own languages to address AI language bias

Unresolved issues

Whether English should continue as the imposed common language or if alternative multilingual approaches should be adopted

How to address the technological limitations of current translation systems, particularly for oral communication

How to develop sufficient data sets for the thousands of underrepresented languages in AI systems

How to balance practical communication needs with linguistic diversity and cultural preservation in international forums

Whether AI will eventually provide a universal solution to language barriers or if human-centered approaches are needed

How to address the political and cultural implications of language dominance in global governance

What the future common language might be as global power structures shift (Chinese Mandarin was suggested as a possibility)

Suggested compromises

Providing multiple language options with simultaneous interpretation in several major languages instead of enforcing one common language

Using language labeling systems to help participants identify shared languages and form communication bridges

Accepting English as a practical solution while acknowledging its imposed nature and working toward more inclusive alternatives

Leveraging AI and technology as culturally neutral tools while building better data sets for underrepresented languages

Allowing cross-linguistic communication where speakers of related languages can communicate in their native tongues while understanding each other

Combining technological solutions with human-centered approaches to address both practical and cultural needs

So what language does AI think in? AI is thinking AI can think in every language All 7,000 languages But if we don’t have enough data sets Then it thinks in English And other major languages… AI actually can talk in Other than the 7,000 languages They can design their own Native computing languages It’s different from human languages

Speaker

Ken Huang

Reason

This comment fundamentally reframes the language barrier discussion by introducing AI as both a potential solution and a new complexity. It reveals that AI has its own linguistic limitations (defaulting to English due to data availability) while also having capabilities beyond human languages through native computing languages.

Impact

This shifted the conversation from purely human-centered language solutions to technological possibilities. It prompted Virginia to immediately recognize this as a potential solution worth exploring further, leading to discussion about visiting their booth and whether ‘AI is our solution’ for future communication.

I almost got a migraine Everyone was speaking whatever I was not understanding I was only picking from the Swahili speakers So those are the African languages… It’s chaotic

Speaker

Audience member

Reason

This brutally honest reaction captures the real human cost of language barriers – the physical and emotional stress of being excluded from communication. It provides visceral evidence of why the Tower of Babel experiment was meaningful.

Impact

This comment grounded the theoretical discussion in lived experience, validating the premise that language barriers create genuine suffering and exclusion. It reinforced the urgency of finding solutions and made the abstract concept of communication challenges tangible.

So I do think English as a common language Does work in an international context English is the solution for the chaotic Tower of Babel situation So I do think English is the solution for the chaos

Speaker

German/Chinese speaking audience member

Reason

Coming from someone whose native language is German but who chose to communicate in Chinese during the experiment, this represents a pragmatic conclusion based on direct experience. It’s particularly insightful because it comes from someone who experienced the chaos firsthand and made a reasoned choice.

Impact

This comment introduced the first strong argument for English as a practical solution, setting up a debate between idealism (multilingual inclusion) and pragmatism (English as lingua franca) that continued throughout the discussion.

And then just coming back to the point of India has so many dialects We have no national language because language is very political… English becomes the default language Because it’s politically the most equal

Speaker

Hindi-speaking audience member

Reason

This comment reveals the deep political dimensions of language choice, showing how English can paradoxically serve as a neutral option when local languages carry political baggage. It also highlights how newer concepts like ‘internet’ may not have native language equivalents.

Impact

This shifted the discussion from viewing English as purely imposed to understanding it as sometimes politically neutral. It introduced the concept of language politics and the practical reality that technical terms often lack native equivalents, adding nuance to the debate.

I want to challenge the notion of one solution of English being one solution… why not give people more options… what if in the discussions that we have there are more simultaneous interpretations in at least a handful of different languages that people have choices

Speaker

Stephanie Borg Psaila

Reason

This comment reframes the entire debate from a binary choice (English vs. native languages) to a multiple-choice solution. Drawing on UN and EU models, it offers a practical middle ground that acknowledges both inclusion needs and practical constraints.

Impact

This comment elevated the discussion from either/or thinking to both/and solutions, introducing the concept of strategic multilingualism. It moved the conversation toward more sophisticated policy solutions rather than simple language dominance.

AI doesn’t think in any language it thinks in terms of maths mathematics the proximity of different notions which are coded in digits… it is mathematics actually all digital or everything which is digital are digits zero or one… let’s trust and explore this world

Speaker

Audience member

Reason

This comment provides a fundamental insight into how AI actually processes language – not linguistically but mathematically. It suggests that digital solutions might be culturally neutral in ways human languages cannot be.

Impact

This deepened the technological discussion by explaining the mathematical foundation of AI language processing. It offered hope for truly neutral communication tools while also prompting Virginia’s humorous observation that binary language only requires learning ‘two letters.’

Almost 95% of data sets or languages data in the internet is English so the rest of languages and the people are not communicating in their own internet language… AI cannot understand your own culture your own native languages or your own native like local like the knowledge

Speaker

Una from China

Reason

This comment exposes the fundamental data bias in AI systems and connects it to cultural preservation. It reveals how technological solutions may perpetuate rather than solve linguistic inequality, while also describing community-based solutions.

Impact

This comment brought the discussion full circle by showing how even technological solutions reflect existing power imbalances. It introduced the concept of community-driven language preservation and highlighted the cultural dimensions of the digital divide.

Overall assessment

These key comments transformed what began as an experimental demonstration into a sophisticated multilayered discussion about language, power, technology, and inclusion. The conversation evolved from experiencing the chaos of multilingual communication to exploring three distinct solution pathways: pragmatic acceptance of English dominance, strategic multilingualism with multiple official languages, and technological solutions through AI. The most impactful comments revealed the political dimensions of language choice, the limitations and possibilities of AI solutions, and the deep connection between language access and cultural preservation. Together, they demonstrated that the language barrier in global governance is not just a communication problem but a complex intersection of politics, technology, culture, and power that requires nuanced, multi-pronged solutions rather than simple universal fixes.

Is AI our solution for multilingual communication? How will we communicate in the future using AI?

Speaker

Virginia (Ginger) Paque

Explanation

This question emerged after Ken Huang presented AI’s capability to think in multiple languages, prompting exploration of whether AI could solve multilingual communication challenges in global governance

What language does AI think in, and how does it handle the 7,000 human languages?

Speaker

Ken Huang

Explanation

This raises important questions about AI’s linguistic capabilities and limitations, particularly regarding data sets and default languages in AI systems

How can we develop better oral communication translation technology beyond current text-to-text translation?

Speaker

Una (from China)

Explanation

Current AI speech recognition can only handle about 100 languages with limited quality, creating a gap in real-time multilingual communication

How can we address the data imbalance where 95% of internet language data is in English?

Speaker

Una (from China)

Explanation

This imbalance affects AI’s ability to understand and process non-English languages and cultures, creating barriers to inclusive communication

Could internationalized domain names provide insights for multilingual internet governance?

Speaker

Virginia (Ginger) Paque

Explanation

Domain names must work across different languages and writing systems, potentially offering lessons for broader multilingual communication solutions

How can we create community-driven projects where people contribute native knowledge in their own languages?

Speaker

Una (from China)

Explanation

This addresses the need for more inclusive data collection and knowledge sharing that preserves cultural and linguistic diversity

What would happen if we provided multiple language options with simultaneous interpretation rather than defaulting to English?

Speaker

Stephanie Borg Psaila

Explanation

This challenges the single-language approach and explores how offering choices in major languages could improve participation in global forums

How can we better utilize cross-language communication where people speak their native language and understand responses in another language?

Speaker

Virginia (Ginger) Paque and Stephanie Borg Psaila

Explanation

This explores the phenomenon of asymmetric multilingual communication as a potential solution to language barriers

Is English an imposed language or simply the most practical solution for international communication?

Speaker

Virginia (Ginger) Paque

Explanation

This addresses the political and practical dimensions of language choice in global governance, questioning whether English dominance is problematic or pragmatic