TidyToolsTranscripts · updated

Video transcript · 19 min · 2,743 words

Keynote on cognitive power and democracy at Wikimedia Futures Lab: transcript

Josef Holý, a Prague-based misinformation expert, opens the Wikimedia Futures Lab with a keynote on cognitive power and democracy. He argues that big tech's extraction of data and attention, frictionless design and monopoly power undermine shared truth. He urges Wikipedia to see itself as public service and critical infrastructure.

Key points

Chapters

  1. 0:00Introduction and reframing the talk
  2. 1:01McLuhan and three angles on technology
  3. 3:00Extraction, attention economy and frictionless design
  4. 6:43Lies, emotions and the post-truth world
  5. 8:57Big tech monopolies and political power
  6. 11:13AI investment bubble and the race with China
  7. 13:17Alternatives: common good and verifiability
  8. 15:28Wikipedia as cognitive power and public infrastructure

Full transcript

0:00So we will now be starting with our opening keynote. I'm very glad to be welcoming Joseph Holy. And Joseph is an expert on misinformation who is based in Prague. He lectures on hacking minds at the Czech Technical University. And he is the co-host of the podcast Canaries in the Net, which is a podcast about algorithmic propaganda. Thank you, Joseph. Hello. thank you eva thanks everyone thanks for the i have my clicker thanks for having me it's a great honor to be here i love wikipedia of course and uh i i pay on a regular basis i have to say so so i'm paying at least trying to to pay my debt because uh because what you guys are doing is of course a great service to the society and i want to get to that as well i have reframed the the name of my presentation and the theme a bit uh from you know shifting landscape in europe

1:01because uh it's not about shifting landscape it's about you know the the rack being pulled out under our feet so to speak so i refrained it to cognitive power and democracy because this is what we are really talking about and this is what we should be focusing on let me explain uh i always or very often not always very often start my presentations with this guy his name is marshal McLuhan. And he was one of the biggest media theorists of 20th century. And his famous quote said, medium is the message, which he wrote a book about. And what it means is that what matters in regards to media is not only what is flowing through the media, but how the medium works and what impact it has on the society.

1:50Mášel Macluhen byl řekl, že vlastně všechny technologií a informáčního mídě, většinou většinou, většinou, jak lidé pracují, jak lidé pracují a většinou. Většinou mídě a technologií většinou struktury většinou. To je to, co jsme představili. můžeme vytvořit v této historii některé třeba. Závětší větší věci jsme představili jsme představili představili, ale jsme představili elektromagnetické mídiy, radice, tv a také.

2:27A všechny věci jsou vědětší věci dramatické věci, jak lidé se strukturence a jak se vědětší. We basically see and we can analyze the impact of the current AI-driven, digitally-driven development on the society from three angles. One is what are the goals of the technology of those who are building the technology? How are they building the technology? What is the nature of the technology?

3:00And then who is the one or who are the ones who are controlling the technology? so as for the goals for the current state what is going on right now in the so-called big tech sphere the big tech arena the main goal is extraction right data is the new oil we all know that this is what even the economist has published like five or six years ago on their on their main page uh shoshana zubov in her influential book the age of surveillance capitalism calls this need for extraction she's calling it extraction imperative meaning it's not a side product of doing other things those companies are not building websites and collecting data on the side they are collecting data as the primary as the primary goal as the primary way of working, so they have to.

3:52Extract data as much as possible, even opportunistically, in order, because it's hedging against the future, the more data you collect today, the better statistical models, meaning artificial intelligence, you will build in the future. This is what Google has realized 20 years ago already, or 25. But it's not only about data, because what they are extracting and harvesting is our attention. This is the ultimate goal. Most of the internet currently runs, apart from subscription, run on advertisement for which you need behavioral data and large scale analysis.

4:26That's the reason why we have those statistical models these days. And this is the reason why they are built by the same companies which gave us social media, because these are the ones who are best positioned to actually build those large language models or generative AI in general, meaning pattern matching and then generating information, generating statistical engines, basically. This is what the attention economy, this is what we are living in. Herbert Simon in 70s identified it as a very quite simple economical relationship between the amount of information and our attention.

5:03If the information grows, then the attention becomes scarce resource, basically. So this is the goal. Extraction. Extraction of data, extraction of attention. Then what is the process? I am calling it here for the purpose of this called frictionless information. We can look at it like that because frictionless design is a paradigm of Silicon Valley. All the product managers in all the teams in all the small or large services and apps, how the way how they are approaching the process of developing a new feature new functionality is through frictionless design so to speak right so they want to remove any friction from user's way so so the user fluently goes through different workflow workflows let it be in the work setting or to you know sharing an app on social sharing a photo on social media and you have a whole methodology around frictionless design you have a whole methodology around addictive design because in order to to grab our attention they need to make their products addictive right we all know

6:07that that's the dopamine feedback loop which social media algorithms have you know implemented so so well and there is again a whole methodology how to build habit forming products those products are addictive and similarly to extraction imperative where extraction of data is not a side product but is a key activity. Addictiveness of the products is not a side product. It's the key design principle, basically. And this basically leads to the problem with truth, which we have right now as a society.

6:43Influential study from MIT from 2018 has shown that lies spread six times faster than truth. Thanks to social media algorithms. It was a bit more complicated, but we can live with this headline. A výsledky výsledky říkáte, že nejvící, že nejvící nejvící nejvící, které nejvící představují, nejvící, nejvící, nejvící, ale nejvící nejvící výsledky výsledky.

7:17Takže výsledky je výsledky výsledky, by algorytmi, protože algorytmi preferují kontent, který zároveň kterým výsledný výsledným výsledním výsledným. To typu kontent, kterým výslepším výsledný vysledným výslepným výsledním výsledným výsledným je výsledným, takže výsledným nebo výsledným And it's not about facts and truth anymore, it's about emotions. So now some people are saying we live in a post-truth world, right?

7:52But the definition from Wikipedia says that it's that basically objective facts matter much less than emotions and personal opinions. This is being taken for granted. Everyone is like, oh, this is inevitable result of development. what I call this complete bullshit, we cannot live with this. If we don't refuse this very strongly, then we are erasing five, six hundred or at least three hundred years of cultural development in Europe.

8:25Our Enlightenment and democratic emergence or re-emergence of democratic societies since 18th century, let's say, is standing on being able to find consensus on truth, to find consensus on shared perception of reality. And if we build our understanding of reality on emotions, then we have a very strong problem, very big problem. We need facts, we need truth. That's why we need Wikipedia, of course. So that was the process, frictionless information, addictive design, and so forth.

8:57And then power, who controls all of that? Who is running it? These guys are running it. Peter Thiel, who, by the way, has been running around the world for the last 20 years, saying that democracy is an outlived concept and it has to be replaced with something else, potentially with something what is called dark, dark enlightenment. But that's for another for another talk. but his famous quote from one of his books says monopoly is a condition of every successful business he said that he will not invest into any startup which doesn't strive or at or you know want to become a monopoly right he says that monopolies escape the trap of competition that monopolies are the successful businesses so they as capitalists want to centralize power just think about it just think about the mindset right so monopoly is the condition of a successful business so now we are in this new paradigm where american big tech monopolies have merged with the

9:59political power so you see this photo of donald trump inauguration from last year where you have all the captains of the big tech industry standing in the first line even in front of the members of the future government. So you have a merge of technology power, largest corporations in the world, merging with the most powerful government in the world, arguably. And then you have a complete shift in power where the president of the United States is sitting in the Oval Office and Elon Musk is giving a press conference there, which has never happened before.

10:41I don't think even Rockefeller, you know, dare to have press conferences from the Oval Office of White House. So this is the new paradigm, right? This is the power. And these are the new monopolists. They are not startups anymore. They are the most valuable companies in the world. And the centralization of the whole industry can be seen on pictures like these, which are circulating around Bloomberg and Financial Times and so forth. The way how the whole AI revolution is funded right now looks like a bit of a Ponzi scheme or circular investment.

11:13Economics, no one knows how this will end, whether they will achieve their goal or whether this will crash somehow. The truth is that most of those companies have a lot of money, so they can probably, you know, uh effort to to lose some in the process but now right now we have not millions or billions but we have hundreds hundreds of billions of dollars being invested Into large data centers and that money don't even change hands often, right?

11:44It's just an accounting operation because NVIDIA gives money to OpenAI and OpenAI uses that money to buy chips from NVIDIA. So you have a mutually beneficial exchange of money. No real value is produced, but everyone is reporting revenues and profits and so forth. now they also are racing right we have AI race big tech is saying we are racing against China we need to beat China so don't regulate us because we are the ones who will help you democratic western societies to win against big bad China I'm not saying that China is small and good China may be big and bad but this is the race which they have started but the question in the race is what is the trophy based on what i have said it's probably monopoly but in their words trophy is uh you know sort of the utopia of agi artificial general intelligence which is directly taken from books like culture series for sculpture culture series from

12:45and banks which most of those guys have read before right The alternative, which we see already happening, is this, 1984. When you merge information, cognitive power with state, and you have a handful of monopolies collaborating with the most powerful state in the world, you are heading to the same model, which they have in China, just in the Western setup. Europe is trying to regulate, but when you are regulating, you are defending, and when you are defending, you are losing.

13:17and our regulations will have hard time being implemented because, of course, American establishment is fighting against them. Donald Trump is threatening by tariffs on a regular basis, right? So what can we do about it? I'm running out of time, so let me go through it very quickly. As opposed to extraction, we should aim for common good. And that's what you guys are doing. This is the quote from Jimmy Wales. Imagine a world where every person on the planet is given free access to the sum of all human knowledge.

13:48That's what we are doing. This is a completely different statement than I want to invest in any monopoly, in monopolies only. Grounded information as opposed to frictionless information. So verifiability as a key principle. You guys have a constitution. your practices and your code of conduct and all of that consists of a constitution of a public infrastructure here. The question only is whether AI could speed up the verification and the verifiability.

14:23And then the third, who owns it? Well, Wikipedia is distributed. We have seen the numbers, how your readership and contributorship is declining, but you still are number seven or number eight in the world. Still better than Pornhub, obviously, which is a good sign. uh but uh the the title of this table says most visited websites in the world but the problem is that these are not websites when you look at the at the list you have google youtube facebook instagram gpt reddit and and then the rest so most of the top most of the top 10 so to speak are not just websites those websites are controlling our perception of reality because they are serving billions of people on a daily basis how the world works how the world looks like what is important what is not and how should they interpret the events and and and other phenomena it can be tied back to you know the famous or well-known plateau scave allegory basically these this guy uh showing the

15:28shadows on on the wall these this is the big tech and we are the people behind the wall not able to see anything else. Because they gained majority in the information space. All the media, not just you, but all the media, all the traditional media are on decline and big tech is rising. So we are not talking about websites. I would like to urge you, stop thinking about Wikipedia as a website. It's not a website. It's a cognitive power center. You are still holding a cognitive power because you are focusing on a verifiable truth and you have all those processes, which you very well know.

16:08and we have a tradition in cognitive power in managing cognitive power and developing it for the good of society europe has a hundred years old tradition of public service media public service media represent agreement and consensus of society on the fact that cognitive technologies information technologies media are way too important to be left to markets and democratic society need and deserve public service media which will basically try to objectify reality through good journalistic practices and resources and so forth. So the question is how can we transfer this from the centralized model which we have been having for the last 80 or 100 years to the new decentralized network world.

17:00Right? So that probably has to go through communities. We cannot afford for the infrastructure to be governed centrally, because otherwise we will head to the 1984, which I described, and community-based, Wikipedia is community-based, basically. So Wikipedia is a public service infrastructure already. You are already doing it. You are relying on a lot of, you know, smaller donations, etc. And the way how you are managed in a very pluralistic and decentralized way is the way how, you know, to know where to go.

17:32So Wikipedia is not just the public service infrastructure, but critical infrastructure. And what I would like to review with is, don't look at yourself as a website, look at yourself as a source of cognitive power and think about how the ways, how your platform works, how you guys work, can be scaled up to other domains of the information ecosystem. Stop defending Wikipedia, stop focusing on představují představují lidí. České lidé musíte představit představit informací, takže webzávají se představit předmětí představit.

18:12Musíte to představit. Představit je, která je, která je, která je, které je, která je vytvořit, která je, který je vytvořit, která je vytvojí, která je vytvoje vytvoření informací ekosystem. Thank you very much for your attention.

Source and licence

Recording: File:Keynote at Wikimedia Futures Lab in Prague.webm by Anna Mendez (WMDE) (2026-04-26), via Wikimedia Commons. Licence: CC BY-SA 4.0. The transcript, summary and chapters on this page are shared under the same licence (CC BY-SA 4.0).

The transcript is machine speech-to-text by the TidyTools Audio & video transcriber; the summary, key points and chapters were written by Claude from that transcript. Both may contain errors: check the recording before quoting it.