OpenAI is officially calling on the UK government to implement a legally binding regulatory framework for advanced artificial intelligence. Tom Duff Gordon, the company's policy lead for EMEA, stated that the current political climate offers a unique window to establish stable, long-term safety requirements for the most powerful AI models.
The proposed framework would shift from voluntary commitments to mandatory enforcement. OpenAI suggests that regulations should focus specifically on top-tier laboratories and models that pose potential systemic, national security, or cybersecurity risks, rather than imposing identical restrictions on startups or general AI applications.
Under the suggested model, the UK AI Safety Institute would take a central role in oversight. Key components of the proposal include mandatory third-party safety testing, rigorous cybersecurity audits, and the implementation of a formal incident reporting system for any significant model anomalies.
This shift in stance aligns with broader industry trends as leading firms move from opposing regulation to participating in its development. OpenAI notes that such a framework would help the UK harmonize its governance approach with emerging standards in the United States and the European Union.
