Europe Eyes ChatGPT Rules
European regulators are weighing whether to classify OpenAI’s ChatGPT as a “very large online platform,” which would impose stricter obligations under the Digital Services Act. ( ) At the same time Brussels is debating a “Digital Omnibus” that could loosen parts of the AI Act and GDPR even as U.S. debates focus on liability — OpenAI backs a bill limiting liability for large-scale AI harms, and analysts are flagging the industry’s “hidden financial loop” and competition risks. ( )
European Union regulators are deciding whether ChatGPT belongs under the bloc’s toughest online-platform rules. (ec.europa.eu) (thehindu.com) The European Commission said it is analyzing whether OpenAI’s service should be designated under the Digital Services Act after OpenAI disclosed that ChatGPT search averaged 120.4 million monthly active recipients in the European Union in the six months ending September 30, 2025. (help.openai.com) (mobileworldlive.com) Under the Digital Services Act, the threshold for a very large online platform or very large online search engine is 45 million monthly users in the European Union, and a designated service gets four months to comply with added duties. Those duties include assessing systemic risks, opening data to vetted researchers, and facing direct European Commission supervision. (ec.europa.eu 1) (ec.europa.eu 2) Brussels is debating that tougher designation while also advancing a separate “Digital Omnibus” package that the Commission published on November 19, 2025 to simplify parts of the Artificial Intelligence Act and other digital laws. The Commission says the package is meant to make implementation “timely, smooth, and proportionate.” (digital-strategy.ec.europa.eu) The two tracks point in opposite directions: one would pull ChatGPT into stricter platform oversight, while the other would trim compliance burdens across the European Union’s broader digital rulebook. Industry lawyers say the Omnibus package now sits in talks between the Council and European Parliament after both adopted negotiating positions in March 2026. (lewissilkin.com) (addleshawgoddard.com) The Digital Services Act is not an artificial-intelligence law. It governs online intermediaries and focuses on illegal content, user rights, transparency, and systemic risks created by very large services. (ec.europa.eu 1) (ec.europa.eu 2) In the United States, the current fight is less about platform designation and more about who pays when artificial intelligence systems cause harm. Wired reported on April 10 that OpenAI backed Illinois Senate Bill 3444, which would limit liability for developers of frontier models in cases of “critical harm” if they did not intentionally or recklessly cause it. (wired.com) (ilga.gov) A bill summary says Senate Bill 3444 would create an Artificial Intelligence Safety Act and shield developers when a frontier model causes critical harms, provided the developer did not act intentionally or recklessly and made required safety disclosures. As of February 4, 2026, the Illinois General Assembly site listed the bill as referred to assignments. (trackbill.com) (ilga.gov) Competition officials are watching a different pressure point: how the artificial-intelligence business is financed. The Federal Trade Commission said on January 17, 2025 that cloud-provider partnerships with OpenAI and Anthropic can create lock-in, deprive startups of key inputs, and expose sensitive information that can weaken competition. (ftc.gov) Tech Policy Press argued on April 13 that some of those deals form a “hidden financial loop,” with cloud credits, booked revenue, and rising valuations reinforcing each other even when underlying demand is still forming. Europe’s ChatGPT decision now lands in that wider argument over whether artificial-intelligence companies should be treated more like neutral tools, mass online platforms, or infrastructure firms with outsized market power. (techpolicy.press) (ftc.gov)
