A Chinese company has been building AI for years to predict who is going to criticize the government before they do so

In ‘Minority Report‘, Tom Cruise was the head of the pre-crime police, a department capable of arresting criminals before they could commit the crime in question, all thanks to the powers of mutants or precogs. Well, according to the New York Timesthere is a Chinese company that is trying to build a similar system, but their target will be future political dissidents and instead of mutants with powers they will use AI. what’s happening. The leak reported in the New York Times contains internal documents from the Chinese company Geedge Networks and has been published by a group of researchers at Vanderbilt University. In it they detail how the company is building an AI system capable of predicting which citizens will become political dissidents in the future. Geedge is investigating how to use LLM to synthesize large packets of data (including browsing histories, locations, online activity and contacts) and then infer citizens’ behavior, detecting whether they will present a “political risk” in the future. Like the police precrime, but for political dissidents. What is Geedge?. In September 2025 we learned that a Chinese company was exporting the surveillance system known as “Great Firewall of China” to other countries. It was Geedge Networks. The company, which has one of the creators of the Chinese firewall as a key investor, has already sold its solution to countries such as Kazakhstan, Pakistan, Ethiopia and Myanmar. What this great firewall does is analyze the traffic of entire countries, even capturing personal data such as passwords and emails. Why is it important. According to the leak, the system is in the research phase, but it is still a disturbing approach. It is no longer just about using AI to monitor what people do, the next thing is to anticipate what they could do and even think. We see every day that AI models have biases and make many errors, using them as predictors to repress dissent poses a terrifying scenario. Tech authoritarianism as a service. As we said, Geedge is already exporting its solutions to other countries so it is selling technological authoritarianism as a service. The worst thing is that we do not find this only in China, but it is a global trend: the United States too you are delegating critical security functions to private and disreputable corporations like Palantir, and The United Kingdom also wants to follow in their footsteps. The bottleneck. There is good news (if you can call it that) and that is that Geedge has encountered a problem in developing this system: they do not have the power to manage such a volume of data. According to the New York Times, since they cannot access the most powerful chips due to the US blockade, since 2024 they have been forced to use AI models and less powerful chips. In order for the system to be able to manage the enormous amount of data they already collect, they need computing capacity that they currently do not have, always according to US sources. Image | Xataka with Gemini In Xataka | We have been hearing for years that China scans the faces of millions of citizens every day. It’s already happening in Europe

We criticize the EU a lot with its obsession with regulating Big Tech. There are at least two examples that justify this obsession

The Digital Markets Act (DMA) and the AI Law They are two of the great exponents of something that the European Union is highly criticized for: his regulatory obsession. It is true that these regulations restrict companies and can slow down European innovation – this has happened with AI – but these worrying side effects are accompanied by others that are much more welcome. Especially because this regulation has made the world a little more interoperable. There are two great examples of this. First example: USB-C. The adoption of the USB-C connector as the mandatory Being able to charge mobile devices and other hardware products is undoubtedly positive for users. Although the standard has its own problemsits use as a universal connector has avoided the use of proprietary connectors that made interoperability difficult and caused greater problems for the environment in the form of electronic waste. Second example: Universal AirDrop. We have also recently seen how Google offered support on the Pixel 10 to be able to transfer data to an iPhone or iPad thanks to AirDrop support in QuickShare. That support will be extended to other Android phones soon, and that improves interoperability between both platforms. From now on it will be much easier to transfer photos directly from mobile to mobile (be it iPhone or Android) wirelessly, and there we have to thank the European Unionwhich forced Apple to modify the way AirDrop works to comply with the DMA. And there is still more. These efforts to improve interoperability will soon be even more rewarded. Google and Apple have announced their collaboration in making portability between different platforms much easier. Thus, changing from an Android mobile to an iPhone or vice versa it’s going to be easier thanks to the efforts that both companies are making. Why have they made that decision? Again, due to the “regulatory obsession” of the EU. The EU sticks out its chest. Euroregulators in fact celebrated this decision by Google and Apple these days, and affirm that the renewed interoperability “is an example of how the Digital Markets Act (DMA) offers benefits to both users and developers.” That same regulation was what allowed iOS 26 to add support to transfer an eSIM to and from an Android mobile, for example. The EU against (almost) everyone. The EU’s regulatory obsession may often be criticized, but the truth is that it is the great reference when it comes to confronting the unlimited ambition of Big Tech. It has done so in the past with the RGPD or with the DSA and the DMAand now with the AI ​​Law. In all of them the ultimate goal is normally reasonable, although it often happens that the regulation ends up being exaggerated or, as with AI, comes too soon. The last chapter of obsession. European regulators suspect that Google is using content from news publishers and other creators to train their generative AI without permission and without offering compensation. These practices may constitute an abuse of Google’s dominant position in the market, which would negatively affect both competition and content publishers themselves. This research also affects “AI Overviews,” which extract and summarize information from other websites, potentially reducing traffic to those original sources. Brussels Effect. The application of these regulations in a market like the European one causes the so-called “Brussels effect”. For large technology companies such as Apple or Google, it is more efficient and profitable to adopt a single standard for all their products worldwide than to design specific versions only for the European market. Thus, this obsession not only benefits us European citizens (when it does), but also ends up becoming the de facto standard worldwide, as has happened with the USB-C connector. This regulation ends up becoming a powerful engine of global change. It is not perfect by any means, and we are seeing it with the AI ​​Law or the cookie nightmare, but even in those cases the EU seems to have realized and is trying to change things. The challenge of the AI ​​Law. If the DMA pursues interoperability, the AI ​​Law seeks transparency and compensation to prevent these monopolies from consolidating in this era of generative AI. The investigation into Google is not only a defense of copyright, but a preventive measure against competition. Meanwhile, the US and China seem turn a blind eye and we have even seen how the leaders of big technology companies They ask that copyright laws not be applied arguing the famous “fair use” of those contents that have little de jusot, at least for content providers. In Xataka | All the big AIs have ignored copyright laws. The amazing thing is that there are still no consequences

Log In

Forgot password?

Forgot password?

Enter your account data and we will send you a link to reset your password.

Your password reset link appears to be invalid or expired.

Log in

Privacy Policy

Add to Collection

No Collections

Here you'll find all collections you've created before.