The following are the outputs of the captioning taken during an IGF intervention. Although it is largely accurate, in some cases it may be incomplete or inaccurate due to inaudible passages or transcription errors. It is posted as an aid, but should not be treated as an authoritative record.
***
>> IHITA GANGAVARAPU: We are here. We have a couple of speakers yet to be set up. We have tech support somewhere. Yeah.
>> IHITA GANGAVARAPU: Good morning. Thank you for coming to our meeting. Our title: Democratising Access to AI. It is a 60‑minute section where we have half of the time for audience interaction. When we talk about and we have start we have a few speakers offline. We do have a few speakers offline, including the online moderator.
When we talk about democratising access to AI, we are making sure that artificial intelligence technologies and resources are accessible to a broad range of people, not just the large corporations or the large‑scale participants. The goal is to make sure everybody is empowered, even the small businesses, educators, researchers, and organisations from all backgrounds and all economies and the benefit from AI.
Development and stimulation of AI, particularly the large language models are dominated by technology companies right now. That space has critical issues around access, control, and input. With the models that are activating, they are risking consolidation of power and limiting the technology. When we speak of open sourcing LLMs, we are looking at creating a pathway to democratise AI.
Potentially, reducing the cost and fostering innovation, by enabling more and more stakeholders to participate in the development of AI. Today's discussion is going to be focusing on the strategic, economic, and social implications. The potential to contract the controls and increase broader distribution of technological benefits.
Before I start, I would like to introduce you to our panelist. I'm joined by Daniele Lee. Online we have Yug Desai who is the moderator. We have Purnima Tiwari who is the rapporteur for the session and a speaker joining us remotely. Thank you for joining us.
We will start off with the discussion. Very first policy question is to Daniele. How does open sourcing influence innovation within the AI industry and what long‑term implications of AI on the structure of the tech industry itself?
>> DANIELE TURRA: Thank you so much for presenting me today.
I'm glad to be here to discuss the very important topic. Everybody right now is talking about AI. Open source has been around for a long time. The narrative has been in a way influenced by the large big tech giants that we had just mentioned. Open source has in a way a different history. Especially free in open source software.
Before getting into the specific, you know, industry implications, I would like to spend a minute to just, you know, introduce once again the cause of the open source. We can just start by saying that open source was a philosophy that was first in a way brought forward by Richard Salomons in the United States. He believed that open source should mean, you know, sharing the code. But they also tightly related with the concept of freedom.
Free not as in here, but free as in freedom. Freedom of speech and especially the four freedoms that define the core ideas of open source. The freedom to use code, the freedom to study code, to redistribute it, and to modify it. There should not be any large or small entitled to in a way own strong Intellectual Property that code and this, of course, is an idea that can benefit so many actors from the smaller to the larger. But in a way, it can enable others to join the industry as well.
When it comes to the AI context, I think I had a few slides. I don't know if tech support can put them on. We're talking about specific solutions and software that are always, you know, created by two different parts. There's not only a general idea of open source software. We're talking about models and weights. There are mainly when you produce a model, there are mainly two files. One is about the weight and one is about the model itself.
Based on this, the open source initiative published the free and open source idea that defines both models and weights to be fully open sourced. Anybody that can access a truly open source LLM has access to both the model and the weights. It is also the result from the training of the data set used to train that model. Then on the other side of the spectrum at the very opposite side, we have the idea of a fully closed model. You know? When you are just maybe accessing it through APIs or anything like that. But it is already in production. It is not something that can be, in a way, you know, inspected, modified, or redistributed; right? The models that have been defined as open weights.
Where, you know, there are licencing where you are as a researcher maybe allowed to explore either the model or the weights again. But we are not really entitled to visit for commercial purposes, and, of course, most often you are not even allowed to, you know, redistribute it itself. This, of course, creates situations in which not everyone can actually benefit from those models. And the ‑‑ again I would like to stress that the only definition that's truly compliant with open source as we free and open source software as we know it is the one that embodies all four freedoms as defined by the thinkers of the free open sources thought. I don't know again if we have lights. I don't know about my timing right now. But again we can think.
Next slide please. Here you can see the framework I was talking about. Different licencing and not all providers actually have the same models and the same licencing for the model that they provide. In this slide, there's AI as a service tech. I would like, you know, to bring the focus again on components that are needed to build an AI solution from end to end. A few scholars, you know, traced some comparisons between the cloud computing model. Each of these layers open source software can be employed. We should also ask ourselves are we actively as private entities or public entities and so on. Are we really entitled to have something that is truly close source, even if we are influencing so, so many community evidence coming from the entire open source community.
In a way, this can be some food for thought. Thinking about the different steps that are implementing actively AI solutions. When it comes to actively industry impacts, we can also think of all of the open source software that goes both at the AI software service training and fine tuning the models down to the actual infrastructure that's needed to compute to have the computing power to have those solutions actively built; right? Because in the end, as the last slide shows it, at least change the slide to the last one.
Last slide please. The last one. There's supply chain that starts from the data collection. The data storage, data preparation, algorithm training, amplification development. Having open source can something that can benefit the private sector.
Again, I would and this is an entire invitation to think in terms of who builds the software and all of the different steps that are needed to get to it and how technologies are already around there can actively help in achieving truly, truly open source models. Thank you.
>> IHITA GANGAVARAPU: Thank you so much for your points.
I would actually want to ask Melissa when she is joining us remotely. Answer the same question. How does open source influence AI innovation in the entire industry and what are the long‑term implications? Over to you.
>> MELISSA MUNOZ SURO: Okay. Can you hear me? Perfect.
Good morning. Everyone. I'm Melissa Munoz. I work here for the government in the Dominican Republic. We are using technology to improve life and make everybody interactions with corporate America more efficient, inclusive, and even more enjoyable. I wanted to answer this question in straight and case of what we're doing here at uptick.
One of the most exciting ways that we're doing this in the national strategy in which a big part of that is – basically, it is an open source AI system that in the future will make the government faster, smarter, and even more personal. That's what we're trying to do. China isn't ready yet. Right now we're focusing on laying the ground work with a project. Open source technology plays a key role in the project, because in a word, it opens the door for more collaboration and innovation. We are building a strong foundation by collecting and organising the data and that we need basically to make it work.
How does it work? We're collecting data from the government systems. These people led the insight for how they interact with the services. We have how to set up specific interaction where people can contribute to the data. How they phrase requests, questions, and this isn't about collecting personal information. It is about understanding the way Dominicans communicate. AI reflects certain culture and language. This is for a collaboration between government, citizens, and universities.
This is basically to ensure the data is accurate. It changes the technology. We are breaking the big tech companies instead relying on their tools. We are creating solutions. They have specific needs. For us that means understanding and Spanish and it is like a culture. That doesn't stop there. Open source means that other Spanish-speaking companies can do. We have the corporation waiting. I'm working closely with AI development. And open source isn't just about influencing innovation rate. It is about basically the shaping how technology serves people. That's why we take in the art.
The innovation at the end shows how open source can empower governments too and also engage citizens and create opportunities. The technology should make life simpler. Ultimately, happier and open source is a key tool to shift and create the inclusive and accessible people‑centered tech industry. Thank you.
>> IHITA GANGAVARAPU: Thank you so much for your points, Melissa. You also highlighted on certain initiatives. I would now think the same question applies.
We would like to take a comment or question from the audience before we move on to the policy question.
>> ABRAHAM FIFI SELBY: All right. Thank you very much. I'm happy to join the session. I'm from the Global South in the African context. I will be speaking about the viewpoints. They have African content. In terms of influence and innovation, we see in Africa that we have a democratisation of AI development. That means there's a very low cost in terms of developing the AI systems. It is very high. We have data. Investment has to go through before we get up.
So, this open source AI systems is helping young people to bring out innovation. Let me go to fostering collaborations. Physically the world is changing. Everything also evolves around technologies now moving into AI. We cannot focus on advanced countries and also not look at the counties. At least the open source systems is helping the people from the way out to make sure. They also provide a source of data which can improve the other regions.
With the other regions, they can platform. Imagine in Africa, there's some countries in the policy document creation. Unless it is business people who needs to get information about which policy, I'm doing this kind of government contract. They are connected to the source and policies across Africa. Between Africa, they can be able to get and ACI and other American countries.
Let's also look as addressing the local needs. This AI tools that we use physically. We have languages that is because of involving in Africa.
Now, collaborating, this AI tool in terms of addressing. It might be the global source. We are talking about this in open source system. This is the standard. We need to look at documentation. Some kind of data to help us in the Global South contest. Let me ask move in to assessing what we need in Africa. There's no improving on the livelihood. There's long‑term implementations that can help the tech industry. I'm looking at the tech industry around the globe and specifically moving into the African contest. There's implementation in Africa. I know you have developed the AI strategy document. They are struggling. It is a very good way for innovation. We see them having the data source about everything and open AI.
For instance, every system is from them. They have the large language models. They have everything. This AI tool is moving. Africa is lacking behind. Why? We feel if the government can develop on their own. We need to connect to our other sources. We need more investment and more collaboration on that. Also to review. They can also do their own AI models. That can also support the development in Africa. It is what our other speakers help about in the country. We are not getting it done. I'll leave my colleagues to talk. Thank you.
>> IHITA GANGAVARAPU: We have talked about accessing the local needs. We want to hand it over to you. If you have a comment or question, you will be having the Q & A round. We will be happy to hear from you. Yes.
Meanwhile if there's an informant or question in the chat or from the online participants, let us know.
>> AUDIENCE: Thank you for this. My name is Leena. I work for research for Common Ground. We work in conflict effectively. We are going to deploy AI with trust and collaboration. There's two challenges that I see. You just mentioned ethics. We're trying to build things on top of the commercial models. There's still an ethical question about where the data goes and how much we are sure whether or not that data is being unused to train those models.
And the second is: you said Africa needs to build their own models. They will dominate the market and end up becoming major revenue builder. There's something a little bit naive about the idea that we can compete. There's no competition unless you regulate the monopolies here. Otherwise there will be no. It is kind of two questions. Thank you.
>> IHITA GANGAVARAPU: Two minutes left to answer the questions.
>> ABRAHAM FIFI SELBY: There's no prerequisite for this. We need global Africa. Despite that, we have to encourage our government to have some investment in infrastructure. What I wanted to address is that our local needs were emphasising that. The Global South cannot come in and the data source that is we need local.
Let's say we have some language models in Africa, Swahili. Let's say Arabic, French, and other stuff. Portuguese. All of these things address in the African country. If we Africans not making to build some contest that can connect to the large open source systems. They will be lacking behind. This is the way the ethics comes in. We must copy from what the global is doing and build upon ourself. We can always not rely on the Global North from the Global South perspective. We must build our own models to improve our AI development and AI policy. This is how I was addressing. You made a clear point.
Based on the investment in ethics and investment in infrastructure, I really agree with you. Africa is lacking. We don't have the infrastructure. We must all rely on the Global North. To rely on the Global North, we must contribute to the Global North perspective by providing data that address the local needs. You can copy the information when you need according to the AI development. Thank you.
>> IHITA GANGAVARAPU: Is it okay if we limit the questions to 15 minutes? We have to move on to the second set of the panel discussion. Then we'll pick it up.
>> AUDIENCE: The question I want to ask is directly related to the monopoly issue.
>> IHITA GANGAVARAPU: Go ahead.
>> AUDIENCE: I'm disagreeing with the last question. I thought the whole point of open source is it was open. Essentially our sharing. If you developed the model, you may not have open source. I understand that Meta is halfway there and maybe full. Maybe the first presenter. The point is you can have access to that. You don't have to have the investment. You can use it.
In regulation, it presumes they are monopolies. You are going to regulate how they ‑‑ I don't know, how they sell it. Which to me does not distribute the knowledge and does not distribute the capability and it is much better to do open source than it is to go through regulation. I'm confused about how you are approaching the issue.
>> IHITA GANGAVARAPU: Our next question is on monopolies. Maybe, Daniele, if you want to keep it under a minute to address this.
We can pick up the discussion in a few minutes.
>> DANIELE TURRA: Sure.
Actually, I was about to introduce some of the points of my help in the sense of the following question. But in a way, I agree on the fact that open source as a tool is a philosophy that can help in not really systemically regulating monopolies, but sharing the knowledge and giving the opportunities to other actors to, you know, again get the skills and not be, you know, blocked by specific Intellectual Properties in that. So this is a way to do it again. And Meta is again as you said halfway there. They have an open weight model. Some are not commercially available. They have researchers to use it.
All in some context, the models also from other providers are allowed to be used in the commercial context. Again, that's always think about the skills and resources needed to build those models. And if some actors are really, you know, should be entitled. Really put a wholly open source definition on that. Because the reason the entire supply chain. I would like to offer some open wash. We need to really categorise things and call them. I hope we could help some of your doubts.
>> IHITA GANGAVARAPU: Can you hear me?
>> YUG DESAI: Yes.
>> IHITA GANGAVARAPU: I'm going to pass the floor to you.
>> YUG DESAI: I would like to pass to the question that we have on innovation.
Hello. Are you there? It is for the landscape and what is necessary to manage this? To answer the question, I want to pass it on to Bianca.
>> BIANCA KREMER: Hi. Everybody hear me? I would like to address ‑‑
(No audio)
>> From your perspective. The first one is on open source. We will have questions about the subject. We have the difference between the close source and open source elements. This is the first question that we will address. The second one is what qualifies as an open source LLM? This is actually really important for us to address the challenges that we will face on this. The third one is which and where are specific cases and corporate case where we can find possibilities with open source LLMs?
After that, I'll bring some experiences from Brazil on open source platforms. In a way, we can address the competition problem as well. Being developed by Universities in our country. It is hard to do this. This is AI models. We have been talking about it. You'll find the training and source code made available to the public. It is not only the developer, but us researchers and organisations from Civil Society as well. We cannot only improve itself, but make it better in a way that companies and business models are not interested in achieving due to economic purposes.
How do they come from the counter sight? They are maintained by companies. It means you cannot access the model. The older open source on the other hand is free to download, modify, and free to be adopted. These products have been instrumental in making it available to the public in order to not only address social problems that we have been facing in the development of these technologies in certain societies. Since we've been talking about the Global South, for example, in Brazil, we had a concrete case that we would like to share with you about a deputy. Which she wrote on ChatGPT. Cavella is a world community. I don't know if you've heard about it before. Black woman in Cavella.
What happened is the image was generated about a black woman holding a gun pointing up. She didn't write anything about guns. It happened. It was a case of what we've been studying the last ten years of what we call racism in platforms. It is a case, comfort case about how the Generative AI technologies have been developing. Not talking about the gender bias that we all know when we have been written two years ago. Name ten philosophers. They were all European and white men. Then you say there are no women. They are always white women and European or North American. Also always white. These are some biases that we have been facing in the use only of these platforms that open source, for example, could be open to address and also to modify the model and addressing the topics of solving some problems of bias. Not all of them. Because when you have algorithm, you always have bias. But some of them. This is something I don't want to talk too much. Just to address the topic of what we've been talking about.
After we have open for questions and things, I would like to also exemplify with the cases of activity and AI. They are two birds from Brazil. We have several birds in the region. They are both birds. These are projects from public universities in Brazil, developing the open source technology. We have been very successful in developing the technologies in Portuguese.
This is the second and third part that I want to have my remarks. I can hear my older colleagues. Just to not only clarify and also make imagery of what we've been talking about. I am from CST. Centre for Science and Technology in Brazil. I'm a member of the Steering Committee, a political position. In the political position, this year we were very successful to have a forum of Internet governance. It was in Africa. A lot of members from the Portuguese community have been discussing the topic of how to build technologies and why not discuss the development of LLM technologies in Portuguese in a broader community. In the broader perspective. It is even more dramatic for us.
When you go to African countries, for example, they speak Creole among them. Much more than Portuguese. This is something we've been talking about. It is a matter of sovereignty. Just to address the topic in the first place, I would like to thank you for the attention, thank you for the opportunity, and I keep myself open for questions and to exchange with my colleagues.
>> YUG DESAI: Thank you, Bianca. Now over to Melissa. Are you able to hear me?
>> MELISSA MUNOZ SURO: It is my turn?
>> IHITA GANGAVARAPU: Yeah. I'll take it over. Since we cannot hear you. I would like to pose the same question to Daniele. When you understand in what ways can the open source models present in localising?
>> DANIELE TURRA: Thank you for the question. I will be honest and critical. I think that technology licencing itself cannot really alone prevent large corporations from taking over in the sense. As I stressed earlier, I believe that we have to in order to protect the actual definition of open source and software.
Of course, software sharing practices and business can help as the ‑‑ one of the men here and all of the audience started to in terms of monopolies. So when we are talking about licencing, I believe that open source is a deal. But also open parameters models can be a good way to achieve that, you know, sharing of knowledge around ecosystem. And be a place in the topic and Global South. Some of this is not open source. But it is still an important role.
Again, I would like to stress about the resources. I think that having, for example, a publicly managed infrastructure could be something that can help us. You know, just like they started to point out this large models of development. The companies with private money. It doesn't mean that we cannot in a way benefit from those altogether. Not all businesses, especially SMEs, can employee this model. The Global South is poor in terms of computing power.
Therefore, it does not have enough power to train this model; right? Fundamentally infrastructure, there's lacking in a sense these makes also in the sharing and management in the infrastructure. In that sense, we could also think in terms of building the models and running the models in production. In general in both cases, I would say that one important thing that I would like to bring as a proposal is to have large cloud businesses that have the computing power offer this capacity for free whether it comes to truly open source models that can be in a way published as open source and true open source.
For example, the location of the computing power could be managed by a law, by, for example, let's call it a computing tax of some sort. Maybe partnership with some Civil Society Organisations that work with the freedom of open source production. That's in terming for example, the Eclipse Foundation. They supervise a lot of efforts in the open source community. It is in the production of open source. There's a lot of work here. I've seen a few and raised a few eyebrows probably.
Again, this is for everyone. If we get a look at how open source works, we can develop the open source model for the future. Thank you.
>> IHITA GANGAVARAPU: Very good answer, actually. I would like to request joining us online for the comments on the question. Keep it under three minutes please. Thank you.
>> MELISSA MUNOZ SURO: Okay. I will do my best. When I was mentioning earlier about registry back in the Dominican Republic.
>> IHITA GANGAVARAPU: There's issues. We can't hear the online participants.
>> MELISSA MUNOZ SURO: Can you hear me well? Hello?
>> YUG DESAI: Seems they cannot.
>> IHITA GANGAVARAPU: Yes. You are audible now.
>> MELISSA MUNOZ SURO: Can you hear me well?
>> IHITA GANGAVARAPU: Yes.
>> MELISSA MUNOZ SURO: Basically, what I was mentioning earlier in the Australia and become back in the Dominican Republic is core principles of sovereignty. This ensures the total systems that we create under the national control or public access and privacy. That will develop in the case to the entirely from scratch. Platforms like, for example, GPT4 are proposed. They have dependency in data. They have a lot of the science systems that align with the national priorities and values, ensuring independence and security and managing the technologies. Open source AI models can produce corporations. They mention the system. It is for the original corporation.
Basically, by using open source which we take control of our tools. And data. We have the presentation ensuring that technology serves for public interest. However, open source is not without challenges. Building and developing the system is access to code. It demands for technical expertise what high quality data in the area where developing countries like mind was focused on fulfilling the implementation.
One of the biggest challenges they are facing right now with open source AI is having the right to make it work. The model needs powerful comparison around what we call GP closers. They have the Dominican Republic that's hard to justify. There's so many other priorities. Health and poverty. They like external services. They already handle for you.
With open source, we are the ones that have to set it up and make sure that's something good to have in mind. They performance how we need them to. Open source doesn't come ready to solve every problem. You have to fine tune them and teach them. That takes time, expertise, and more money that we don't have in the Global South.
As one of them was speaking before, there's also the data. We have in the governance system is messy and scattered. It is not always useful for a project. For example, we have to work hard to clean it up the data and combine it with new information from different government platforms. We have also set up places where citizens can share how they talk and ask questions, so the AI can build and understand the cultural and language.
Finally, the cost of keeping everything running. Open source sounds great. Because you are not paying someone else every time you use it. It is the systems working on the long term. You have to operate hard work and make sure you can handle more users and the realities of the open source AI isn't something that you just throw out and forget about it. It needs investment, planning, teamwork, and that's why we're working for other countries and try to make it regional. And international. We can share resources and make open source AI the solution, of course, for everyone. That's it. On my part. Thank you.
>> IHITA GANGAVARAPU: The points were important. Now back here for the short time, we would like to open the floor for all of you for interaction. What do you see open source and such as the potential for misuse or incentives or investments in AI and research. How can they be by the development and harnessing the opportunities. The floor is yours. Do we have anyone that would like to add a comment or question?
>> AUDIENCE: Thank you very much for your panel and the interesting discussion that you were having. It is not an answer to the question, but more a question to you. I want to ask you more concretely about what are the enablers of open source? LLM and open source AI. We were touching on the competition issue. We see thus far from the market incentive; Facebook goes some extent. They are lacking to open the model only to fully reuse and just fill out the slides and big tank companies. There's no real incentive to do it.
Their models might have the ability to outperform open source models for a time. We're talking about the issues of, like, a smaller languages than English. I guess there would need to be some kind of, like, common data sets and open data sets. They were touching on the issue of how to distribute computing power. I was wondering, maybe, if you could make concrete recommendations on what would we need to build and how could we do so to set up a system in which open source LLM and open source AI can thrive. Thank you very much.
>> IHITA GANGAVARAPU: That's a very detailed question. Could we take one more comment or question before we let the panelist answer that? Do we have anything from the online participants?
>> YUG DESAI: How do we ensure cultural nuances from Africa as well as AI sovereignty? That's all I have online.
>> IHITA GANGAVARAPU: Perfect. The second question did answer quite a bit. We can have any comments on the first question please for the panel.
>> DANIELE TURRA: Yeah. I'll try to be very brief.
One key difference that we can see in open LLMs when it comes to, you know, their nature. It is the product. Without specific computing power needed. When it comes to training, we need GPU infrastructure. That doesn't come cheap. It is a way to the communities of languages. I might have brought a few proposals earlier. The general takeaway message of folks here to have is this. Let's try to redistribute and better share that computing power.
>> ABRAHAM FIFI SELBY: Okay. I'm asking about data source in Africa. We are now growing in terms of the digital landscape and economy. Data sovereignty is something we cannot leave out. It is conducting and related to the policies. I understand that we can see in terms of governance and structures. There must be government, academics, and all of the stakeholders must come together to develop the policies.
There's a risk it might be done. They can touch upon it. There's also the private entities that provied risk office. We must involve everyone. The government cannot do this alone. They make academia. We cannot do the policies without them. There's the correlation they were built to address and address the open source. Thank you.
>> IHITA GANGAVARAPU: Perfect. Thank you so much. We are close to the closing. I would request the panelist to spend 35 seconds to give us maybe one critical area that, you know, we should all be focusing on within the open source LLM.
>> DANIELE TURRA: I will not reiterate my message. I made it clear a couple of times. I would like to say sometimes they challenge the idea that make some of us not happy. Give us some concrete ideas on how to tackle the global issue. It is better to start with something that's not perfect and pivot and adjust from it. Rather than just trying to make everybody happy. That's it. Thank you.
>> IHITA GANGAVARAPU: I think this was a good start.
>> ABRAHAM FIFI SELBY: Building, there must be an established mechanism. That can help with the private and public partnerships.
Also, the investment that supports the open AI and AI development within the Global South. By connecting researchers and innovators to the advance counties that are global. Which can also bring back to the Global South development. It is what I'll say. It is a very useful section that I really appreciate and expect. We cannot build that together. Thank you very much.
>> BIANCA KREMER: Thank you.
Very briefly, I'm less optimistic. I do believe that to address competition and the advance of the economics, we need regulation. I'm from law. Maybe I'm bias.
In Brazil, we have just discussed the AI bill and protection laws that last six years. I believe that when we have regulations of the topics and the participation of government on the development and the free development and industrialisation or deindustrialisation of our countries for our participation in economy. If we don't rely on regulation, regulation on the topics. We won't move forward in our own development. Not only as country, but also as economic partners.
In the Global South, for example. This is why I'm not that optimistic. I do believe we need more enforcement in terms of legal participation in the process.
>> IHITA GANGAVARAPU: Thank you, Bianca. Melissa, if you can hear us, you have the closing remark please.
>> MELISSA MUNOZ SURO: Can you hear me well?
Well, I think we should focus on trust and collaboration. This means prioritising data and being transparent about where the data comes from and how it is used and ensuring it is really, really good at the end. With universities and research and development that truly represents in the case of governments, for example.
Also, an invitation to invest in the collaboration and make AI accessible. Ethical and inclusive AI should be at center of our first, ensuring that technology works for people, builds trust, and attracts sustainable investment at the end. That's my final thought.
>> IHITA GANGAVARAPU: Thank you. We have come to the end of the session. Thank you for joining us.
When we talk about democratising access to AI, the spectrum of concerns that come in to play for highlighted.
Thank you for joining us. I hope you carry forward the deliberations and future recommendations in the IGF and after. Thank you.