ChatGPT will be subject to reinforced rules within the EU in terms of user protection, provided for by the European legislative arsenal, as will the online gaming platform Roblox, very popular with children, and the social network Reddit, the European Commission announced on Monday.
These services have thus joined the list of major online platforms and search engines subject to this regime, as provided for in the European Regulation on Digital Services (DSA), after having all three crossed the threshold of 45 million users within the EU. The services concerned by this status must in particular evaluate and mitigate the risks incurred by their users, such as exposure to illegal or dangerous content.
ChatGPT thus becomes the first AI (artificial intelligence) service that will have to comply with this more restrictive framework. Responsible companies risk heavy fines (up to 6% of their annual global turnover), or even a ban on operating in Europe in the event of serious and repeated violations.
“This will set a precedent”
Like Roblox and Reddit, ChatGPT will have four months, i.e. by the end of November, to comply with these reinforced obligations. The ultra-popular AI model from the American giant OpenAI declared that it had more than 159 million users in the EU at the end of March, or more than a third of the European population.
Roblox, a vast American online gaming platform targeting children, with its immersive universe with cubic and colorful graphics, narrowly exceeds the threshold provided by the DSA, set at 45 million users (or approximately 10% of the inhabitants of the entire EU): it claims 46.6 million active players in Europe. For its part, Reddit, a large online discussion forum also based in the United States, on which one can discuss almost any subject, says it has more than 57 million “redditors” in the EU.
The inclusion of ChatGPT in this list, as a “very large online search engine”, could constitute a major development, believes Lena-Maria Böswald, interface expert, an NGO based in Berlin. “This will create a precedent,” she warned ahead of this designation, which had been expected for several months.
Numerous investigations launched
“OpenAI was already subject to risk management obligations due to the European law on artificial intelligence (the AI Act, Editor’s note), but the DSA could broaden the scope beyond the risks linked to the model itself, by extending it to the operation of its conversational robot, to transparency in the use of data, and to its impact on users and society in general,” she judges.
The DSA has allowed the EU to launch numerous investigations over the past two years targeting large online platforms, which resulted in the announcement of sanctions against three of them, Elon Musk’s social network X and the e-commerce sites Temu and AliExpress.
