Anthropic has spent weeks warning that AI could end human civilization. All the while, it was building a wet lab. The company only confirmed the wet lab after reporters started asking questions.
Eric Kauderer-Abrams, Anthropic's head of life sciences, confirmed the lab in a Reuters interview and said some experiments will be done in-house and others through outside partners.
“We believe that to do biology, the final test is still, and will be for a while, in real lab work,” Kauderer-Abrams said. “We absolutely are doing that today.”
An Anthropic spokesperson told TechCrunch that the lab is focused on fundamental biology, not specifically drug discovery. Anthropic has separately said it wants to do preclinical work on rare diseases and other areas traditional drugmakers do not find attractive financially.
Amodei said last week that AI could cure most major diseases in five to 10 years. That is an extraordinary claim from the same CEO who has been warning that AI poses an existential threat to humanity.
Anthropic has been spending on this for a while. It bought Coefficient Bio in April in a stock deal reportedly worth about $400 million. Novartis CEO Vas Narasimhan joined the board. Claude Science now taps more than 60 databases covering genomics, proteomics and structural biology. This month Anthropic added a Life Sciences Verification Program for vetted researchers.
Anthropic also wants Claude on the machinery. The company says it is training the model to direct robotic lab systems with limited human intervention, including protein and nucleic acid characterization. It released a Model Hardware Standard in August. Kauderer-Abrams called the work “very early innings” and said humans remain involved for safety.
Anthropic says it is not doing coronavirus or gain-of-function work. Fine. But the context here matters. In June, then-DNI Tulsi Gabbard released declassified material linking U.S.-funded research at the Wuhan Institute of Virology to the pandemic. The FBI and Department of Energy both lean toward a lab leak. Nobody is accusing Anthropic of anything. But a secretive AI company running biology experiments with limited public oversight is exactly the kind of thing the public has earned the right to watch closely.
Reuters also reported that Anthropic had already found examples where its own systems could have been used toward biological-weapons development.
Researcher Jacob Coxon resigned after warning that people building AI “earnestly believe that it could kill us all by the end of the decade.” Anthropic's alignment lead put the odds of AI wiping out humanity within 10 years above 10 percent. Amodei called for a slowdown last weekend and again pointed to bioterrorism.
Investor Chamath Palihapitiya responded after the wet lab became public: “The group behind such hits as ‘We’re All Going To Die’ and ‘Regulate Me Now’ are building a wet lab in SF. I do not recommend this.”
Novo Nordisk, Bristol Myers Squibb and Roche's Genentech already work with Anthropic, which raises an obvious question about who benefits from what gets discovered. Stephanie Walter of HyperFRAME Research put it directly: "That gives Anthropic a direct feedback loop and potentially valuable proprietary data for improving Claude." Its pharmaceutical partners will want to know what Anthropic keeps, how their discoveries get used, and where the company draws the line. So far those answers have not been forthcoming.
Kauderer-Abrams said Anthropic will not run human clinical trials for now. Its work will stay in early preclinical research, including areas drugmakers have avoided.
"We're not competing with pharma and biotech companies that make their business in bringing drugs to market," he said. Life sciences already ranks among Anthropic's biggest investments by headcount and resources. Job postings call for lab procurement, biochemical expertise, and automation. One listing says the goal is to accelerate life-sciences research "by an order of magnitude."
They warned you AI could kill us all. Then they hired a biochemist. Draw your own conclusions.