That's because Italy's data protection agency announced on Friday a blockage of ChatGPT within the country's borders. Italy has also started an investigation. Called a "temporary limitation" in the English translation of the press release, the ban arrives on the grounds that ChatGPT does not adequately protect user data or have any kind of mechanism to prevent minors from using the generative AI service.
Under Europe's General Data Protection Regulation (or GDPR), ChatGPT parent company OpenAI has 20 days to comply with the order, or else it could face fines of up to 4 percent of global revenue or €20 million, whichever is larger.
There's a whole host of interesting angles to this. For one, it's hard to argue with Italy's complaints. A recent bug exposed private user chat titles, and OpenAI later admitted the breach was larger than originally thought. There is also a complete lack of gatekeeping for minors.
Beyond that, though, is the issue of "training" the AI itself. ChatGPT and other AI chat-bots like it basically teach themselves how to act like humans by reading things humans write on the internet, without the consent of said humans. As TechCrunch explained, ChatGPT is able to recite information about individuals while also regularly getting details wrong. There does not appear to be a way for the individuals in question to correct these errors.
