Anthropic says its AI biology lab used Claude to discover a novel enzyme system with CRISPR-like repeats, but has deliberately kept humans supervising the work.
Anthropic reported that its biology lab has already produced a significant finding: a novel enzyme system featuring CRISPR-like repeats, discovered with Claude's help. The company was explicit that it has not let Claude operate the lab autonomously — humans remain in the loop on every step.
The disclosure doubles as a research and governance statement, showing a concrete scientific result while drawing a clear line around autonomy limits in a high-stakes domain.
AI labs are increasingly using scientific discovery as proof of commercial value beyond chatbots, and biology is the highest-leverage, highest-risk arena to prove it in. Keeping humans in the loop is a deliberate risk-management signal to regulators and enterprise biotech partners, not just a technical choice.
The daily signal, curated. Get it in your inbox.
Subscribe on LinkedIn →