Government finally draws a line on unchecked AI power

Kanishka Narayan's phrasing — "must not be allowed to 'let rip'" — is unusually blunt for a government statement. It signals that the era of self-regulation is being treated as failed, not merely incomplete.

The biosecurity lab backed by public money and contributions from OpenAI and Anthropic is a concrete step. It moves safety work out of corporate blogs and into a facility with government oversight. That matters because biosecurity risks are tangible, measurable, and harder to dismiss as speculative.

Operators who build on top of foundation models should watch this closely. When a government funds a lab with the very companies whose models you depend on, the resulting standards will shape what those models are allowed to do — and what you're liable for downstream.

The UK approach contrasts with the US reliance on voluntary commitments. If the lab produces binding red lines, any company deploying in the UK will need to demonstrate compliance, not just publish a safety page.

— Edgar Melo Costa, founder

Founder’s notes are opinion. They are not investment, legal or professional advice.