OpenAI Quietly Lobbies EU to Weaken Groundbreaking AI Act
Despite public calls for global AI regulation, OpenAI secretly pressures EU officials to soften the AI Act and avoid strict rules for its general-purpose systems.
OpenAI CEO Sam Altman publicly advocates for global AI regulation during his recent world tour, but behind the scenes, the company actively lobbies to weaken the EU's landmark AI Act. Documents obtained by TIME reveal that OpenAI pressures European officials to reduce the regulatory burden on its general-purpose AI systems, arguing for amendments that make their way into the final legislative text.
In a previously unpublished seven-page white paper sent to EU officials in 2022, OpenAI insists that models like GPT-3 and DALL-E 2 do not qualify as "high risk" systems. The company claims these tools only become high risk when developers apply them to specific dangerous use cases, an argument that aligns perfectly with the lobbying efforts of major backers like Microsoft and Google.
This covert push to soften the rules directly contradicts Altman's recent public warnings that OpenAI might cease European operations if the AI Act proves too difficult to follow. The European Parliament approves the legislation in mid-June, and it now moves to a final round of negotiations before formal adoption as early as January.