AI Giant Anthropic Dives into Biology: What's Cooking in Their New Lab?

Anthropic has confirmed the establishment of a "wet biology lab" in the Bay Area, enabling its AI models to conduct physical experiments. This development supports the prevalent belief among AI leaders, including Anthropic's CEO Dario Amodei, who recently opined that AI could potentially cure most major diseases within the next 5–10 years. To achieve such ambitious goals, it is understood that large language models (LLMs) would necessitate real-world testing capabilities for their theories. Eric Kauderer-Abrams, Anthropic’s head of life sciences, emphasized this point to Reuters, stating, "We believe that to do biology, the final test is still, and will be for a while, in real lab work." He further confirmed that the company is actively engaged in this today, with the lab operating much like a typical biotech facility, conducting internal research and collaborating with external partners.

The news of Anthropic's wet lab is perhaps unsurprising given its acquisition of Coefficient Bio, a stealth AI biotech firm, in April. While Anthropic chose not to disclose specific projects underway in the lab, it did specify that the main focus is on fundamental biology, rather than direct drug discovery. This distinction is crucial for Anthropic, as it aims to avoid the appearance of competing with the pharmaceutical industry, a sector where it has numerous significant customers and partners. This includes a recently announced deal to collaborate with Novo Nordisk on joint drug discovery. Anthropic has previously encountered backlash for launching products perceived to rival those of its customers.

To further support its engagement with the scientific community and clarify its intentions, Anthropic also launched a Life Sciences Verification Program. This initiative grants vetted bio researchers access to the company's most powerful AI models. Additionally, Anthropic has published recent reports detailing its research efforts designed to support drug discovery, specifically highlighting work on accelerating protein design and uplifting bimolecular modeling.

However, the existence of this wet lab presents a stark juxtaposition against the backdrop of grave warnings regarding AI's potential dangers, many of which emanate from within Anthropic itself. The tech industry is still grappling with the resignation of Anthropic researcher Jacob Coxon, who publicly cautioned that "the people building AI earnestly believe that it could kill us all by the end of the decade." Furthermore, Anthropic’s own alignment lead has previously estimated the odds of AI exterminating humanity within the next decade to be greater than 10%. The escalating intensity of these concerns led CEO Dario Amodei to publish a post advocating for the industry to slow down and implement self-regulation, repeatedly identifying bioterrorism as one of AI’s most significant risks.

This complex and seemingly contradictory situation—where a company deeply concerned about AI's existential threats is simultaneously building advanced biological experimentation facilities—has certainly not gone unnoticed. Investor and AI coding startup founder Chamath Palihapitiya, for instance, commented on X with a touch of irony: "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."