in the next hour with professor Jovan Kurbalija, Diplo Foundation’s founder and director, we will try to envisage what will happen in the AI and digital domain in 2026. As we all know, the year started with many developments and updates. Some of them are shocking and the world is changing faster than ever before.

And it’s changing as we have this webinar and as we talk. So we want to know what will happen with the technological domain? Will it remain isolated from wider developments?

And Jovan, we all want to know what is your take?

Hi Su Sonia, hi everybody. Well, a quick take is no, a tech domain cannot be isolated. It has been the last few days, in spite of huge geopolitical shifts around us, it cannot be isolated.

On the contrary, it will be even more profoundly impacted by geopolitical, geoeconomic, and other shifts than any other domain, because it impacts society, it impacts knowledge, education, works, and other domains.

Therefore, we should fasten our seatbelts. There will be quite a dynamic ride ahead of us in the tech domain.

I agree 100%. Every day, looking at the news, and we also, as you all know, all our participants know, we gather them at the Digital Watch Observatory on a daily basis, and it’s almost impossible to keep up, especially when it comes to AI.

So Jovan, you will be guiding us through the forecast by focusing on 10 developments, I mean 10 subjects, and it’s technology, geostrategy, governance, security, human rights, economy, standards, content, development, and the environment.

So this is a great mapping. Could you please start with outlining the main developments as related to these 10 points, Jovan?

Sure. Those 10 points have been a structure of our prediction sessions and forecasting for the last definitely six years, but we can add some other aspects that are missing, and probably, Susonja, what is important to mention is that this should be an interactive session in terms of the comments asked by participants, suggestions, what is missing.

But let’s start with technology, the first, and I would say the most profound impact, which is in the basis of these changes. There are a few strategic developments of relevance for our discussion. First one, AI is becoming commodity.

What does it mean? You go to the shop and you take AI. It’s not as trivial as it is, but it is rather easy.

And there was one interesting development which went under the radar, which illustrates this point. Apple basically started slowing down their attempts to develop Apple LLM. And their point is, hey, it is commodity, we have thousands of large language models, let’s focus on something else.

Let’s focus on the development around AI. And this will be probably the first major shift. So far we have been bombarded by the news about the latest LLM, how many sort of parameters and the other.

This is basically framing of discussion, which has its logic because of investment and the other things. This is the first aspect. AI is becoming commodity, impacting all layers from the top layer to the countries in battle for AI sovereignty to us as citizens.

Second point, which is extremely important, it is bottom-up AI. We are going to face in during 2026 small large language models that can run definitely on your corporate or institutional servers as we are doing at Diplo. But maybe I would bet even on your mobile.

Now the key question won’t be about having your personal AI, but the key question will be is this personal AI on your mobile connected to some central system, whether it is OpenAI, Cloudy or DeepSeek, or it remains as your local AI completely controllable by you.

This is the critical, I would say, issue for 2026. Does it matter? If it is connected to some central system, like most of the Internet of Things devices are Fitbits, you will be basically passing your knowledge to somebody else.

If it is stored on your system, you will be saving knowledge for yourself, for your organization, Ministry of Foreign Affairs, and the other actors. Bottom-up AI is technically feasible, financially affordable, and ethically desirable. And at Diplo, we are trying to walk the talk.

What I’m just saying now, we are experimenting in Diplo and developing our in-house AI. Therefore, AI is becoming commodity, bottom-up AI, and open source development. That was the critical development in 2025, especially coming with the deep-seek and Chinese wave of open source, reaction by US administration in August, pushing open source very high on the US AI agenda, European Union prioritizing open source together with the quest for sovereignty.

And that will be probably, as we’ll see later on, one of the key geostrategical battles. You probably noticed that I didn’t mention agentic AI and everything that you can hear in media these days. In my view, it won’t be the critical development.

AI as a commodity, bottom-up AI, and open source AI are three things to watch carefully on technological side.

Thank you for that insight, Jovan, and I think especially your take on the open versus closed models and how certain countries are leading the race. I think we’ll hear a lot more about it in this new year and the coming months. And yes, there is another feature, by the way, about our webinar.

Diplo’s AI, Diplo AI will follow these predictions as the year unfolds. And you can write in the chat any specific question that you would like to be monitored. For example, will the AI bubble burst or anything that you’re curious about regarding AI?

And our Diplo AI will be following up through that. And just to make another reminder of earlier what Jovan said, our lovely colleague, Andrijana Gavrilovic, she’s in the chat. She’s our chat moderator.

Please feel free to add your questions and comments in the chat and we will make sure to address them as the webinar unfolds. So on technology, there’s an interesting experiment. If you have a hundred or a hundred thousand or a hundred million dollars, where would you invest it out of five options?

Jovan, would you invest it in hardware, software, knowledge and data, human resources, organization or governance? And while you place your bet, Jovan, I’m curious about you using more so the word knowledge instead of data.

Yeah, but regarding to this bet on out of hundred which our colleagues will be now taking, there is AI Pareto paradox. Pareto is famous Italian sociologist, bit philosopher, who argued that 80% of our investment brings the 20% of real impact and 20% of investment bring 80% of impact. This is happening today with AI.

80% of investment in technology, in hardware, in media cards is bringing just the marginal improvements of AI. Why we are ignoring other four aspects that you can find their software, but in particular knowledge and data, skills and organizational changes. And you’re right, I highlighted knowledge.

The reason is that we speak a lot about data protection and it’s understandable, but data is just ingredients. What is happening currently is that our knowledge is captured. Let us take this example.

If Zoom or I think somebody activated some AI summarizer, are basically capturing our knowledge from our discussion today, it’s not just data. These are insights, reflections, some possible wisdom. And this is what we have to revisit our terminology.

I think data is important, it’s ingredient, but critical battle is for our knowledge. Our knowledge developing events like this, in reflections, family charts, sessions, and other ways. Therefore, yes, I would argue that knowledge should become keyword for the governance discussion in 2026.

Thank you so much for that, Jovan. And now let me use the privilege of being moderator and just insert another question here. Since like in 2021, the metaverse was the hyped word and in 2022, it was blockchain.

And now we’re at the beginning of 2026 and almost no one mentions or remembers these words. So what will happen with the AI lingo and the AI hype with these two keywords related to 2025? Like in 2025, the keyword was AI slop, as most of us have heard about it, which means low quality AI generated content.

And another word was vibe coding, using natural language to generate code. Jovan, do you think they’ll share the same destiny as the metaverse and blockchain?

The answer is yes and no. I think AI hype, not only in economical hype, but also language and cognitive hype, presence of AI issues in public. domain is definitely something which will have to slow down.

No, I would say AI is different than the metaverse and blockchain, because it’s less technological. It’s more societal, especially because of larger language model and automation of the language as a key communication tools of humanity. Now, it’s moving to video and sound, but I would say language is this.

Therefore, yes, the hype has to slow down. But I guess that we will have still AI-related terminology quite present in the public discourse. That’s short answer.

OK, and if I may add, what are your takes on the AI bubble and it’s bursting or it’s not bursting?

Well, AI bubble exists, and it is inflated on an unprecedented level. On the cognitive level, presence of AI narrative in public discourse in media, A. B, on investment, market capitalization on NVIDIA is more than combined GDP of African continent.

I think something like 4.6 billion. And three is a huge investment in data centers and other facilities, which cannot be justified due to the conceptual and mathematical and technical limitations of large language models. They’re hitting the platform of the efficiency.

You can add more NVIDIA cards, but you won’t get better inference and better answers. Therefore, we are hitting the platform. It’s not getting substantively better, but narrative is pumping.

The bubble is pumping. The bubble exists. Is it going to burst?

My argument is no, because it’s too big to burst. Because it can bring the global economy down. Huge money from investment pension fund are invested in companies like NVIDIA.

There will be many small bursts. And I would argue that probably there will be central. decentralization of AI, start-ups, small companies will be basically hoovered by a few big companies and they will survive by using the energy of these small companies.

Therefore yes, the bubble is huge. Is it going to burst like a dot-com bubble or a 2008 bubble? I don’t think so, but there will be small bursts, unfortunately, leading to the further centralization of AI industry in the heads of the big companies that need energy and food to survive.

And actually, that brings me to one of the 10 subjects that we earlier mentioned, which is geo-strategy. Would you like to explain, what do you mean when you say geo-strategy?

It’s an umbrella term, but I would suggest there will be three aspects, basically geopolitics, geo-economics and geo-emotions. Geopolitics is what we are seeing these days, is position of big countries, competition mainly between US and China in tech domain, but I would say EU as important actor. There are some good and bad news.

Good news is that in spite of all geopolitical pressure, we still have integrated global internet, at least on technical level, on the ICANN level. You can still send emails and you can still access different domains. But digital fragmentation is accelerating and it will be the clear dynamics, most likely around two or three hubs, definitely US, definitely China, and we’ll see if Europe will be third hub or not.

That’s probably one of the critical issue for the future of digitalization. We have fragmentation of filtering social media, other services and the other developments. Therefore, watch for that.

ICANN’s future, we have to admit that ICANN can be issued subpoena and ask to remove, for example, domain from its list of domain names. It’s California incorporated organization. That’s reality.

Yes, ICANN is globally in its management, but it is legally part of the California legal system and US legal system. Therefore, that’s not anymore unthinkable that the country can be removed from domain name system. I don’t think it will happen.

I’m more optimistic because it would hurt also US actors, but it’s a possibility. We should also in geopolitics follow the EU sovereignty movement in Europe, kicked off by EU digital sovereignty declaration adopted in Berlin on 18th November. And in particular, latest move at the beginning of the year where EU is pushing for open source and linking open source and sovereignty.

The message is basically, we want to have our own AI and we will develop it through open source and through that, we will achieve digital and AI sovereignty. Another element and hint that it is moving in this direction is that latest Mistral version is basically developed on the deep-seek methodology. It’s open source.

It doesn’t have to do anything with China in a sense, except that deep-seek is a Chinese company, but there are an interesting shifts on the sovereignty level. Second part is geoeconomics and it will be critical in this period because we are seeing that global companies have the huge footprint worldwide. Many of them are based in the United States, some in China, a few in Europe.

And any fragmentation, trade fragmentation, taxation fragmentation will affect them and it will come inevitably. But my bet for today is that geo-emotions will be the most important. If you look at the latest Ipsos survey, how people perceive AI, you will see the least enthusiastic, among the least enthusiastic population is the US population in the United States.

38% of people are enthusiastic about AI. The Asia is much more enthusiastic, I think China with 80% of people who are happy to use AI. Why is this important?

Basically what is happening in the US because of the centralization of AI industry is critical for the future AI. You have very low enthusiasm of the population, caused by different reasons, including the doomsday scenarios the last two or three years. You have very enthusiastic tech industry, who are accelerating, they are basically just moving more and more money.

And you have the administration, US administration, which has a close link with Silicon Valley tech industry, and which is benefiting from the boom in AI. There are some estimates between 35% and 50% of the raise of GDP in the United States in 2025 was caused by investment into AI data centers and other things. Now as we have election year in the United States, you have on one side population which is very skeptical.

You have very enthusiastic Silicon Valley with a strong influence on the President Trump’s administration. Therefore, there will be an interesting dynamics between these three aspects, how it will work. We saw the first hint, where the few states, I think, Desantis and California, push back against centralized AI regulation by the government.

on the level of the United States. They said, no, no, no.

Federal level.

We want to have it locally. Therefore, I think geo-emotions will be extremely interesting, especially in the United States in this election year.

Yeah, and maybe perhaps it’s not exactly what you call geo-emotions, but we’re also seeing a shift to where more and more AI products are becoming about human emotions and about intimacy, and more people are using it as a friend, as a therapist, as a partner, and more products are now being developed.

So I just wanted to add in that sense that content-wise and tool-wise, it seems like from the attention and data economy, maybe things are now more and more targeting our emotions as humans. What would you think about that?

No, definitely. I address this issue more from governance, diplomacy, policy level, but the question of capturing not only data, not only knowledge, but our emotions, or tacit knowledge, is becoming the next step, and that’s basically, I would say, very worrisome development, and with many, many negative impacts.

In this case, I basically refer of the emotions of the U.S. population in this case, very skeptical emotions, and completely different attitude of the economic and political elite. This is going to be, I would say, probably not decisive, but important issue for elections in November this year.

Definitely. Okay, so that brings us to governance. So what should we expect in the new year when it comes to digital and AI governance?

Okay, now, when we speak about governance, I’m coming from internet governance community, and very often for us, you know how it is with person, with the hammer, everything looks like a nail. Now I can spend. and ours discussing IGF, the conclusion of the World Summit on Information, society and other issues.

But ultimate governance dilemma is, our representatives on national level in parliament or on global level, our diplomats, can they protect our interests, our digital interests for our data, knowledge, cybersecurity and other issues.

Unfortunately, what I’m seeing is that we have deterioration of these developments. We have less and less mechanism that somebody who based his business on YouTube, which happened with, I know the case in Serbia, and his YouTube channel is removed, whom he should complain? YouTube ignored him, his government cannot help him, not only Serbian government, I would say, any government except US and Chinese governments.

And we basically have not advanced into this direction. We may have sometimes good events, good discussions, but basically at the bottom, we don’t have any major developments. There are some good news, inclusive governance is still at least notionally preserved through multi-stakeholder participation, although it has many questions to be addressed about this, but we should always ask ourselves the simple question, are our governance efforts helping citizen, company, community in our countries addressing their digital concerns?

Cybersecurity, data protection, knowledge protection, education, IPR. And here I’m afraid that we don’t have a good news, and I don’t expect many good news and developments in 2026. We’ll have many events, discussions, but ultimately we won’t be advancing.

And actually that reminds me like other actors in the… you have mentioned this term techno-feudalism. So when we’re talking about governance and like you mentioned, you’re coming from internet governance where it was all about multi-stakeholder approach.

And it seems to be like this is changing drastically and very fast. What do you think about the term techno-feudalism and are we actually moving there?

Well, I would be even more, let’s say push things even further. I think we are facing the risk of techno-slavery. Feudalism, at least the peasants who were the feudal places had some sort of agency.

If we pass all of our knowledge to big system, there could be new forms of the slavery because what is the slavery? It’s not only physical slavery, it’s basically not having the right to make a choice, not having respect for your dignity and other issues. I won’t be alarmist, but we should be careful about it.

Especially slavery as a method is unfortunately not only not disappearing, but increasing. According to latest statistic, we have approximately 60 million people in some form of slavery, more traditional sex workers, migrants, and other things. I hope that we won’t have attack slavery as a new term entering the digital domain, especially through the question of centralization of our knowledge, personal, institutional, country, community knowledge.

Therefore, yes, the techno-feudalism is likely to develop around fragmentation of the internet, but I’m afraid that we may have go even further in history by having some sort of techno-slavery.

Okay, I will ask a bit further about that, because we’re gonna have other related topics like human rights, so we can go a bit further in that then. But before human rights, we have security. What are your predictions and your highlights about security in the new year?

Again, big domain. We had at the end of the year adoption of Hanoi Cybercrime Convention, which I think is a good development. There is still good discussion at the UN.

In spite of all criticism of the UN, there are movements, there are discussions. I don’t know if these discussions are making us more secure. Like previous example, our governance mechanism are providing us with more tools.

I don’t know. But I would suggest that we put on the radar, in addition to the traditional issues, the question of interconnectivity through APIs with AI tools. We have millions and millions of connections where you connect to ChargPT, to DeepSeek, to Anthropic Cloud and other things, and you create your own AI agents.

All of these tools that are basically connecting are potential backdoors for the cybersecurity attacks. Diplo has close to, I just counted vaguely, close to 500 API connections. Please do not attack us, but it’s a huge vulnerability.

Can you imagine if somebody cuts all API addresses and links through Zapier and other tools? That would be major collapse of the modern digital industry. In addition to cyber war, cyber terrorism, and classical issues, cyber crime that we’ve been addressing, I would put this new risks of new interdependence of AI system through use of APIs.

Okay, thank you so much for that. And I don’t think we’re hearing enough about this, but I think we will.

Completely under the radar, I don’t know why, but it’s completely under the radar.

Yeah, but maybe people or companies… and everyone is avoiding it because now this is the hype, especially with agentic AI as well. And if we put the spotlight on the huge security issue about it, then maybe they’re afraid of things slowing down.

I don’t know. I’m just speculating. And next up, very related to everything that we discussed is human rights.

So we just talked about digital serfdom, digital slavery. And even recently, like, well, maybe we’ll talk about that in content. Anyway, I’ll just give the floor back to you.

What do you think about human rights in the new years?

Well, we have general deterioration of human rights, especially after the change of administration in January last year with Trump administration. Tech companies dismantled content policies often related to human rights. And we know these developments.

And here, the major tension is, is basically between Europe and the United States. Traditional tension was with some countries which were skeptical about human rights. In the Middle East, typically it was also China often, and the West.

Now we have a tension between European Union and the United States on human rights. I won’t go, we can go into specific human rights, but I will go to one issue, which is again, similar to API, not only about human rights, but not on the radar. We have traditional narratives on bias ethics and other issues.

It’s basically probably inflated discussion. But we are less focused on human rights that defines us as humans, our cognition, our knowledge, our dignity, through education, through jobs. And I think that, let’s say, group of human rights is completely, completely, I would say, ignored at least in public debates.

There is some sort of… where you go to AI events and you’d speak about bias and ethics, but what about our right to knowledge, to education? What about our right to remain human?

What about the right to be humanly imperfect, not to be optimized? The human rights community, I would say, has to wake up and see what’s going on around us and don’t follow just inertia about repeating the speeches of three or four years ago. Human rights landscape is changing very fast, and it has a completely new type of human rights, which are ultimately related to three points that defines us as humans.

Protection of our life, physical life, protection of our dignity, and realization of our potentials, intellectual, physical, whatever. And I think we have to get to basics of it. Through this discussion, we will revisit also the core pillar of enlightenment, because what we are discussing today is not just what is the move from Washington or Brussels or Beijing and the tactical, we are revisiting the core pillars of enlightenment, which put the human in the center of societal, economic, and political systems.

And human rights community has to move very fast to revisit that aspect. Okay, and that actually brings me to the next headline, which is economy, and I think it’s very related because it seems like we’re at that point where if it’s good, if it’s good for the economy, then we don’t need to think about human rights.

If there’s a new policy by a tech company and it’s bringing in money, so if it’s beating the economy, we don’t have to think about human rights. But okay, now, about the economy, what do you think and what do you foresee, Jovan? So, Sonia, in tradition…

Additionally, the UN system and the regional organization, you had the three pillars. You had security and safety, you had development and economy, and you have human rights. That triangle, which you can find in almost in the European Union and any organization, completely shifted towards economy and security.

Human rights dimension is basically shrinking. And the key development, and I think my colleague Marilia Maciel wrote a lot about it, is basically shift towards economic security. Economical issues and tech issues in particular are becoming part of security.

Access to rare materials, access to the large language models, and all other issues. And I think this is going to be the major shift in 2026. It started already over the last two years, but it will accelerate.

Tech issues will become security issues. And economy, digital economy, will be centered around it. We have to follow one counterintuitive development.

Everybody expected the fast raise of Bitcoin because Trump administration is very Bitcoin-friendly. What we saw is basically small development, even decline of the rate of Bitcoin in the last year. We should follow taxation and basically existence of the global trade system, and how that collapse of potential, collapse of the global trade system will affect tech industry, especially from the United States.

And this brings us to the next point, which is standards, and which is getting more and more complicated. As you already mentioned, with different countries taking different stances, and this risking decentralization and all of that. what would you have to say about standards in the new year and in this new AI-dominated ecosystem, global ecosystem?

The key issue is to learn something from the past. We as humans, sometimes we are learning, but very often we don’t learn from the lessons from the past. What’s happened with social media platforms?

We were basically got stuck within the specific platform, whether it is initially Facebook, Instagram, Twitter, X, and the other things. We are going to face the same issue with AI. If you have, if we don’t have interoperable standards, if there are no interoperable standards, if you develop your AI system, your weights, your data, your knowledge graph, I won’t go in the terminology, technical, we’ll stay with one platform and you will be stuck.

If you don’t develop it locally, it’s open source. Therefore, interoperable AI standards will become critical for the developments of AI. And by doing that, by pushing in that direction, we should avoid mistake that we had with development of social media platforms.

When you’re stuck, now with Digital Service Act, European Union has been trying to overcome that. I don’t know if it is going to work, but we should learn the lesson and say, okay, let’s make interoperable AI platforms. First advice would be, let’s make your own AI, your company, your personal, but if you cannot do it, let’s make at least interoperable platform where you can move from OpenAI to Cloudy, to DeepSeek, to other platforms.

Okay, and I have to mention the Diplo AI apprenticeship at this point. I’m sorry, it wasn’t in our script, but we’re always talking about this and we also personally to learn about it and the best way to learn. something is by doing and for anyone who doesn’t know or hasn’t heard about it, this is exactly what Diplo’s AI Apprenticeship does, and it allows you to build your own AIs.

But correct me if I’m wrong, are they interoperable, Jovan?

No, we cannot impose the standards on OpenAI and Cloud, but they are interoperable within Diplo system, yes. You can share the things. Thank you, Su Sonia, for mentioning this.

AI Apprenticeship is a very innovative approach where you learn about AI by developing AI. The course starts by developing new AI in five minutes. This is time that you need to develop AI agent.

But you need then at least two months, and I would say much longer to develop good AI. AI which won’t be sloppy like most of the platform, but which will be adjusted to your needs. I think that is very interesting approach.

We basically use the Swiss approach of apprenticeship, which is very deep into the Swiss educational system, to develop approach of learning about AI by developing AI. In this format and this title or some other title, that will be critical for any organization, school, company, ministry department, to raise the awareness of their stuff by developing AI.

I agree 100 percent. And this brings us to content. And I mean, there’s so many, sometimes even scandalous headlines when it comes to AI and content, and the latest being these new futures, for example, with Grok, the X-Platforms AI agent.

Undressing people and people using this, and this touches a lot of things, but. When you say content, I realize it’s a very, very wide topic. Almost anything, everything falls under it.

But what’s your perspective and lens when it comes to content in the new year?

Sure, I’m sure we can spend hours speaking about deep fake AI and other things. But the key issue here is the relation between the United States and the United States tech companies and government and the European Union. We already saw with the fine against tax in the European Union.

That’s, I would say, the critical issue. Are companies going to be, tech companies, forced to obey the law of the European Union? Here we have very simplified narrative, usually in mainstream media.

Oh, European Union, European countries are conservative and other things. And I’m very critical about developments, tech developments and regulatory in European Union on other aspects. But the bottom line is that the country or group of countries can adopt the rules according to their interests.

Who are the tech companies to tell, let’s say, Germany not to have the content policy given the very delicate history of Germany and basically the question of misuse of certain material, especially Nazi memorabilia and other things.

Who is the tech companies to tell 80 million German citizens who voted for government, who have the regulation, that they are backwards because they are trying to filter the content? That’s a very tricky discussion. We as citizens agree on certain laws by our representatives and the other things.

Now you can find in media misuse, censorship and the other things. We have to be very careful to avoid situation of not seeing the forest for the tree. We have always to ask bottom up, bottom question, is particular country, I mentioned example of Germany or EU.

allowed to make rules for the content on their territory. I think they are, and I think tech companies have to adapt to it. That will be the major tension.

There is also Australian rule that banning AI for the younger, the 16, there are many small micro developments, but ultimate basically issue will be tension between European governments, parliaments, and US tech companies.

And there’s much there to discover, but we’re talking about technology and digital stuff. So we won’t get into that between AI and, I mean, not AI, EU and US, but anyway. So we’re reaching towards the end of our list, but we’ll continue, of course, and we’ll take some questions from our audience.

But next up is development.

I have, here’s one good news. So far I have many bad news. It is a good news.

I think that countries which are a bit backwards in inverted commas in AI fields are not necessarily losing. Why? Because I think that the current model of AI developments with basically focusing on hardware, invest heavily into hardware, have the data farms.

If you don’t have your AI model, you’re lost. I don’t think it’s the critical issue. The critical issue is basically how to preserve your knowledge, not data, your local knowledge locally.

And I would say if I were the prime minister of Ruritania, small country somewhere in Asia, Africa, Europe, I would go for that. I would invest more money into schools, into preserving knowledge, creating open source platforms locally. Therefore, that’s what was typical when you go to major AI.

and including UN events, we said, oh, we don’t have LLM. I said, who cares? Even Apple said we don’t care about having a large language model.

There are now close, I think, close to thousands LLMs. But what is missing is how you preserve your knowledge. Therefore, paradoxically, slow entry of the developing countries in the field of AI could be, in this case, a blessing of getting it right.

Investing in what matters. Knowledge, local knowledge, capacity of the people, skills, and different governance.

Okay. Thank you for that. And finally, some good news, at least.

But I’m afraid, okay, the last topic, maybe there’s not too much to say in a positive manner, but I’m curious to hear what you say. But before that, I will do another plug-in right here. Because in our Diplo Foundation’s YouTube channel, we have different series and different videos about all these topics.

And Jovan has his videos, too. And we have these series called AI Shorts. And this was something, the upcoming topic was something that we covered a lot in AI Shorts.

Because it’s surfacing more and more. And it is the environment. Environment and sustainability and how it overlaps with, of course, technology, but in particular, AI.

So, what will we be seeing? Because I’ve heard, I don’t know if that was doable or if it was in theory. But for example, Jovan, I heard in Finland, they were using the heat from servers and AI farms to heat residents and buildings during winter.

But of course, when it comes to this stuff, it’s all about scalability. And Scandinavia and Finland, it’s not that big. And so, there are people all over the world trying to come up with solutions.

And also, there are people who are just completely ignoring this part about AI. the water consumption and all of that, but what should we expect?

Okay, so Sonia, again like with the previous issues, we should step back and ask the questions, do we need these huge farms? What for? And I always start from myself and from Diplo.

I did analysis during the Christmas time, I had some free time, of typical vocabulary that I use and typical framing of issues that I use in my blogs, in texts, in emails and other things. And I came, my vocabulary is not particularly rich, but I came to something three to four thousand words and what really shocked me, I came to about 30 ways how I cognitively frame issues. Zoom in, zoom out, you know my phrase is looking out of the box and these things.

And you really think, and I’m not super creative, but I’m not also a person who follows the crowd, when you really think, we usually maybe use 30, 40, maximum 50 cognitive toolkits. Do I need super trooper AI platform which companies are developing when I can have my small AI basically being extension of my thinking? Or Diplo’s AI being extension of our courses and training?

This is the first common sense question. Do we really need this crazy development of platforms which are using enormous amount of energy and water? I don’t think so.

And this is basically my point. But if we do that, we can use this like finish, we can use it, other approaches to make it cleaner. Since we cannot stop this investment into the big data farms, let’s do at least alleviate the problem and prepare for the moment where we may have huge digital graveyards, once technology, software and anything else.

bypass the need for such huge firms. That’s basically how I would approach it.

Okay, so we’re going to, well, we’re not wrapping up yet. I mean, we still have like around 15 minutes. Would you like to go to the chat?

Would you like to answer some questions that I’ll read out to you? Okay, so why do you think agentic AI is not a highlight this year?

Okay, agents will be used and developed, but they are not going to be as big thing as we were told for a few reasons. One reason is that they are not anymore directly related to technology, having more LLMs and these things. They’re related for human adoption of the AI tools.

And when you have to change humans’ organization, it takes much more time. That’s the first point. And you can see that companies are now, OpenAI is establishing consultancy company, Antropic, with basically dealing with it, why humans are not using AI, existing AI, ChargePT and other things, enough in the nation.

I think that is basically hype, but it has to maintain investment. But it won’t be revolutionarily developed. There will be some good system.

MCP was one of the standards, was a big thing. I don’t think that it will be as revolutionary as we were said. It will be some development, some good applications.

We are experimenting a diploid, agentic AI, but it won’t be the solution for all problems of humanity, as we are told by the media, especially tech media.

Yeah, the media tends to feed the hype. And another question is, well, this time related to diplomacy. What do you think the impact of AI is on diplomacy and peace building?

Okay, on diplomacy per se, impact will be, I would say, probably profound for a very simple reason. Diplomacy is text-based profession. Anything you do in diplomacy, you ultimately have to write the text.

Whether you meet somebody in the corridors of the UN in Geneva, you have to write cable back to your capital. If the meeting is fine, you have reports and meetings. Therefore, that aspect will be profound.

Good news is that diplomacy is more arts than science. Therefore, the question of negotiation, representation, and other human-centered tools will get more relevance. Therefore, generally, I think there are good news for diplomacy and peacemaking.

First one, it will be the end of bureaucratic diplomacy or writing reports. And it will give a chance and time to return to what really matters in diplomacy. Negotiate with the other side, persuade, engage, find a compromise, and deliver some sort of peaceful resolution.

Therefore, I would say, generally, bad news for current diplomacy, which is mainly bureaucratic diplomacy, good news for real diplomacy.

Okay, thank you for that. At least some good news. And there’s like two more questions, but I want to combine them.

Do you think if either of them will, which one will take over? Like, do you think super AI or singularity is hypothetical? Or is there a real possibility of it replacing us or our intelligence?

And another question was, could an AI private ally undermine our wishes? So I’m seeing these two questions as, in the future, do you think AI will take over? or the AI private elite, or neither will happen?

I would say that that discussion is interesting. I love philosophy and I love discussing it, but I think this is not important discussion about singularity and superintelligence. It is like when you have a baby crying and you basically play with some sort of a distraction or what once John Lennon said, and I think that’s probably one of my 30, 35 typical quotes is that life what’s happened to you while you’re busy planning your life.

Now, yes, there might be superintelligence, I don’t know, and I’m a bit skeptical, but okay, there could be. But what is happening now is that things are changing now. Students are using ChargePT to write their essays.

They’re not anymore developing critical thinking. Jobs are lost, very often not for AI, but AI is used as a good excuse. Programmers are basically their job, and there is a bit of cynical developments.

People who created AI are basically the first to feel the basically the automation of their jobs. Therefore, my point is whether there will be an extinction risk or not, I don’t know. We should be very careful to deal with immediate issues which are happening now, and then to keep on the radar extinction risk or superintelligence, but not in the way how it’s presented to us in the global media.

And for me, the key issue is not if AI is going to govern us or rule us. The main issue is if we are going to be ruled by people who own and develop AI. That’s a different thing.

They have a powerful tool in the hands of you, or AI is some sort of abstract development.

And so I’m hearing that, yeah, you’re saying the fact that AI elite will probably, there’s a more possibility of them ruling, being the new ruling class or ruling us even against our wishes. And I personally agree. And then do you think we have a chance, I mean, to at least slow this down or reverse this because like you mentioned critical thinking and I was thinking that maybe AI doesn’t need to become super intelligent if we become dumb enough by handing our thinking over to technology?

That’s I think critical point you’re saying. We say AI will replace us. Yes, if we write useless thesis or useless policy reports, it won’t be that difficult to replace us.

And that’s the point. To whom we compare AI? Generally speaking, many, many texts in the current, I mean, involving global governance are not something really with a breakthrough.

And that’s a good issue that are we going to, on the contrary, I recently read the report on AI from some organization, I won’t mention the negative, I don’t want to mention people in negative context, 120 pages.

And even they didn’t remove the prompts. And it’s a policy document of that organization. And I said, are they expect from me to read 120 pages?

If I’m diplomats in that capital, I would have to read maybe in page 56, there is something dangerous for my country. There were, there are serious problems with us in using, in using it. Now what is the solution?

Why I’m most optimistic, more optimistic, we can develop our own AI. It’s technically feasible, it’s financially affordable, technically feasible, open source, technically, we can afford it. And it is ethically desirable.

And then if I want to use my AI, which is completely my AI, not directly. directed from the big systems, I can decide if I want to automate my life, I can decide if I don’t want any more free will. That’s my personal decision.

We keep it anchored into very simple idea of enlightenment, not only enlightenment, also key religions last 23, 24 centuries, is that human is central player on the earth. And why not with AI? That’s fine, but it’s our decision then.

Yes, I know that. Okay, since we’re almost running out of time before I ask you to wrap up, there’s two more really good questions. Will, do you think geopolitical tensions will threaten energy supplies and how would this impact AI developments and uptake?

Which as we already discussed, like requires large amounts of energy. And we’re having all these AI summits everywhere, small ones, big ones. What do you think about them?

Are they just talk shops or are they helping move AI policy in the right direction?

Well, I’m avoiding them, this big AI meetings for, you know how it is in any policy space. Once the rhetoric starts, you people repeat the same things, ethics, bias, ethics, bias, and the other now energy and these things. We have to ask ultimate question.

And this is this step back. Is AI going to basically increase our three aspects, protection of life, protection of our dignity and the realization of our potentials? This is the first question.

Second question, are our governance AI summits and other things are going to help us to protect our interests on the global level beyond our national borders? Are our governments delivering of the basic social contract? Social contract is very simple.

You pay taxes, male go and serve now female. military service. In exchange, they protect your interests, protect your life, government, protect your life, market economy, political system, and they cannot protect us in digital and the area.

I always ask myself these two questions. And unfortunately, I don’t get positive answers after finishing all of these big meetings. You know, you are sometimes you meet the friends, I am in this community for a few decades.

I like inclusive approach, multi-stakeholder and other things, but those things are rarely answering these two questions. And my other question is how to prevent knowledge seldom or even knowledge slavery. So far, I’m not positive on our results as policy community.

And about the geopolitical tensions around energy and supplies affecting AI.

Yeah, there is now, you can read it everywhere that US has a problem with energy supplies, China is more solid with it because also China is developing smaller models. I don’t know, I don’t have, I can now basically repeat what you can read in the media, but I’m a bit skeptical about this big server farms and use of huge energy and that’s something. But basic narrative is US is lagging behind, they’re opening new, even nuclear stations.

China invested more in the coal energy and it’s now better placed because also of the small models that China is developing, smaller models than US, that’s the common narrative.

Okay, and okay, there’s one, I said it was last, but it’s not because it’s about Geneva and it’s about Diplo and I think it’s a good way to wrap it up before I ask for your final remarks. Do you see international Geneva, including Diplo, emerging as a a true hub for tech diplomacy by 2026, 2027, or are we facing our limits? Is Geneva facing its limits to growth?

Well, Geneva has its, it’s limited by geopolitical developments, obviously. If multilateral system is going to suffer, Geneva will suffer like Vienna, like other places. That’s one thing, but to be a bit of stoic, you may influence these things or may not, including what we just heard this morning, that U.S.

will be withdrawing support, financial support from quite a few organizations in Geneva, which is not good news for multilateralism and I would say global politics in general. But while we cannot impact that, we can do something ourselves. And these are a few things.

One is, and there are very concrete proposals, why not to start something which is in the basis of Swiss educational system, which is apprenticeship? Train the people in international organizations in diplomacy, prepare the global population for the AI changes. That’s the one thing.

Second thing, which is very specific. If you analyze SEO, search engine optimization, what’s going on? People, at least we are preparing reports from Diplo, people are basically visiting Diplo website for verification of information.

Before AI, they were visiting system to get information. They now get ChargePT or Google and then they click on the link. Now, Geneva organizations should become verificators of the global information because somehow there is authority of international organizations behind the reports that they create, documents, declarations.

That’s an interesting strategic shift that simply websites of international organizations should become protectors of the verifiers of the global information system. And third point is the space for Geneva is in standards. It’s all that Geneva is playing an important role, and it has to push that more into AI and other standards.

Those would be three concrete issues that Geneva can contribute. I would advise, and at least we are doing it, we are not repeating something that you can hear at thousands of AI meetings and summits. Ethics, bias, ethics, bias, robots, and other things.

I think Geneva, and Switzerland, and Europe, and any country should find its own specific contribution to AI discussion. And in Geneva is the question of apprenticeship, question of verification of information, and standards.

OK, thank you so much for that. Yeah, we’ve hit our one hour mark. And I know we don’t usually do this in our webinars, but if you’re watching this on YouTube, please make sure to check out our entire channel.

We have master classes. We have experts explain videos. We have things that focus on digital diplomacy and AI updates.

And they’re usually weekly. And I’m also pasting the link in the chat right now. And I’m wishing everyone a Happy New Year and just handing the mic one last time to Jovan to see if he has any closing remarks.

Thank you, Su Sonia. Thank you for joining us today. I hope we raised some issue.

The times are critical. We have to think out of the box, critically, with the common sense. This is basically what we need in order to, I would say, even survive.

Now, deep or small contribution to this discussion will be threefold. First, you will receive the documents with the survey of what I said today in the follow-up to the session. Second, our digital watch will basically monitor questions around this on the monthly, we may say even on weekly basis, for time on monthly basis.

And please send the questions that you would like to be. monitored, in addition to questions that you asked today. And the third point is that AI apprenticeship method will be even stronger on the agenda of this learning about AI by doing AI.

This is critical for governance, diplomacy, and other spaces. Those would be three messages from Diplo. And once again, thank you, Su Sonia.

Thank you, other colleagues, for joining us today.

Thank you, Jovan. And we have all the links that we mentioned in the chat box. And you can also always visit our website.

And the next cohort of the AI apprenticeship starts in April. So everyone has time to check it out and see the requirements. And I’ve met many graduates of it.

And they were all very happy. So it’s a personal recommendation as well. Thank you, everyone.

Thank you. Ciao. Ciao.