CREDIT: CHIP SOMODEVILLA VIA GETTY IMAGES
AI
Sam Altman, CEO,
OpenAI
As a result, the code tackles three main areas that have been on AI researchers and leaders minds:
First, transparency requirements mean companies must document how their models work and what data they used for training.
Second, copyright compliance rules force companies to explain how they obtained their training material.
The third requirement covers“ GPAI with Systemic Risk” or( GPAISR) – essentially the most powerful models including OpenAI’ s o3 and Google’ s Gemini 2.5 Pro that face additional scrutiny including regular safety evaluations and risk assessments.
The penalties are designed to hurt even the biggest companies. Fines can reach € 35m( US $ 38.5m) or 7 % of global annual turnover.
For context, 7 % of Meta’ s 2023 revenue would be roughly US $ 8.4bn.
OpenAI and Microsoft choosing compliance over confrontation OpenAI’ s decision to sign the code is a sizable strategic bet.
The company behind ChatGPT has essentially decided that playing ball with European regulators is worth the compliance costs and restrictions.“ Signing the code reflects our commitment to providing capable, accessible and secure AI models for Europeans to fully participate in the economic and societal benefits of the Intelligence Age,” the company says in its announcement.
Additionally, Anthropic, the AI company that prides itself on safety, has also announced its intention to sign.
technologymagazine. com 151