OpenAI backs mandatory UK rules for frontier AI. The company sets out its position in parliamentary evidence published October 8; the document itself is dated October 5.
Statutory duties should reflect capabilities and risks, focusing on developers of the most advanced systems.
OpenAI prioritizes cyber capabilities, biological risks and loss of control, alongside independent scrutiny and international coordination over serious incidents.
It also reports stronger monitoring and a mandatory alignment check before broader internal Astra use. These are company statements, not independently measured protection.