Sam Altman warns OpenAI could pull out of Europe over AI Act

OpenAI CEO Sam Altman just issued a stark warning to European regulators: play nice with AI rules, or ChatGPT might leave the building entirely.

Speaking at University College London during his European tour, Altman stated that OpenAI could "cease operating" in the European Union if the upcoming EU AI Act imposes requirements the company cannot technically fulfill.

The "high-risk" showdown

The tension comes down to a specific classification in the EU's draft legislation. As currently written, large language models like GPT-4 and ChatGPT could be designated as "high-risk" systems, subjecting them to strict safety audits and burdensome compliance mandates.

OpenAI has consistently argued that general-purpose models shouldn't automatically carry a high-risk label. Altman noted the legislation isn't inherently flawed, but emphasized that subtle details matter and technical limits exist.

Beyond regulations, Altman floated other controversial topics during his talk. He argued that social networks, not AI models, are primarily responsible for the viral spread of election disinformation. He also teased that OpenAI plans to publicly wade into wealth redistribution debates next year once its five-year study on universal basic income concludes.

Outside the hall, tensions were equally palpable. Protesters gathered with signs reading "Stop the suicide AGI race," confronting Altman over AI safety risks before he stepped out to briefly debate them directly.

Why it matters

Pulling out of the EU would be a massive blow to European founders, developers, and knowledge workers who rely on OpenAI's state-of-the-art APIs. Losing native access would instantly widen the productivity and innovation gap between Europe and the rest of the world.

For the AI landscape, Altman’s warning is pure hardball lobbying. The EU AI Act is poised to become the world’s benchmark for artificial intelligence regulation. If OpenAI folds on European terms, other global jurisdictions will likely adopt the same stringent framework.

In our view, actually abandoning the EU market would be near-suicidal for OpenAI's commercial scale. This looks much less like a genuine exit plan and much more like a high-stakes leverage play, designed to water down final technical definitions before the ink on the legislation dries.