Anthropic Just Built a Real Biology Lab to Test Claude
AI that warns about danger is now running real science experiments — here's why that matters.
Anthropic, the company that makes the AI chatbot Claude, has opened a physical biology lab in the San Francisco Bay Area. Until now, its life-sciences work lived mostly on computers — modeling proteins, analyzing data, predicting how biology behaves. Now the company can run actual experiments with real lab equipment and see whether Claude's suggestions hold up in the messy real world. Anthropic's head of life sciences, Eric Kauderer-Abrams, confirmed the lab to Reuters, though the company declined to say exactly what it's working on.
The bigger ambition is automation. Anthropic wants Claude to eventually direct robotic lab equipment to run experiments with very little human involvement. The company launched something called the Model Hardware Standard in August — a set of rules that helps AI systems talk to lab machines. Kauderer-Abrams says this is "very early innings." Think of it like teaching a smart assistant to not just write a recipe, but actually cook the meal and taste the result. The payoff: less time between a promising idea and a real test.
This isn't a side hobby. Life sciences is already one of Anthropic's biggest areas of investment. The company bought an AI biotech startup called Coefficient Bio, and Novartis CEO Vas Narasimhan sits on its board. It's also hiring for lab operations roles. The stated goal includes tackling diseases that big drugmakers ignore because they aren't profitable — a potentially huge win for patients with rare or neglected conditions.
But there's an awkward tension. Anthropic has spent years warning about the risks of powerful AI, including the danger that it could help someone make a biological weapon. Now the same company is building systems that connect AI to real lab tools and real biological material. Anthropic says human oversight still matters and restricts some capabilities. Still, when AI moves from writing text to guiding physical experiments, the stakes get higher — and the distance between a mistake and a real-world consequence shrinks.
- Anthropic (the maker of Claude) has opened a real biology lab in the Bay Area, moving beyond computer-only research.
- The company wants AI to eventually direct robots that run experiments with limited human help — a step toward faster medical discoveries.
- Anthropic warns about AI's biological risks while building AI for biological work, a tension it says it manages with human oversight.
Why It Matters
Faster, cheaper biology research could mean new treatments sooner — but also new safety risks worth watching.