Anthropic says its AI-powered biology lab has already found something worth talking about. The company, which operates a wet lab in the Bay Area, says one of its models found a new enzyme system with properties that look a lot like CRISPR. The enzyme system was found in the DNA of bacteriophages, viruses that infect and replicate within bacteria. It can perform operations like cutting, copying, and pasting DNA.
The discovery was made in 21 hours of concerted effort by the model Claude. Claude searched through data using about 950 agents that burned through 210 million tokens.
What Claude Found
The new enzyme system was found mostly, though not entirely, by Claude, according to Anthropic CEO Dario Amodei. He wrote that a team from Stanford previously discovered a system that is in some ways similar to the one Claude found. Amodei’s framing leaves the question of which system came first as an open one for the research community.
The lab was established this spring, though Anthropic declined to be more specific about how many months it has been open.
The Lab, and Who Runs It
Anthropic published a statement saying its lab “looks like a typical molecular biology lab” doing “research that involves only the lower-levels of the biosafety risk level (BSL-1 and BSL-2).” It added that “All of the lab work is performed by human scientists.”
Amodei said it may eventually be possible for Claude itself to safely perform experiments by autonomously controlling lab equipment, but “we aren’t doing that today.” That is a notable qualification from a CEO who has said one of the things he fears most is that AI could be used for bioterrorism.
“We aren’t doing that today.”
The Bigger Picture on AI and Biology
Amodei has also said AI will “cure most diseases in 5-10 years.” The two statements sit side by side: one about what the company has already found, the other about what it believes AI can do for medicine.
What This Means for Science
The company is making a clear argument about where AI fits in the scientific process. Claude did the search work. Human scientists did the wet lab work. That division of labor is the whole point of the announcement.
The company is positioning AI as a partner in discovery, not a replacement for it. The statement that all lab work is performed by human scientists is the clearest sign of how Anthropic sees this playing out in the near term.
Where the Discovery Fits In
AI-powered biological research is not new. Researchers from Stanford have published papers on work they are doing with large language models and CRISPR. Researchers at UC San Francisco have used AI to design enzymes from scratch. Google launched its first AI-powered biology research tool, AlphaFold, back in 2020.
Yet, Amodei hasn’t ruled out a fully automated biology lab in the future. “Eventually it may even be possible for Claude itself to safely perform the experiments by autonomously controlling lab equipment, with appropriate safeguards in place, but we aren’t doing that today,” he said.
The discovery is real. The lab is real. The human hands are real. The AI that found it is also real.
Anthropic is showing what the future of science might look like: a model finds something, a human team confirms it, and the world learns something new.
Where the paper stands
The paper backs narrow rules against hiding safety failures, such as requiring companies to disclose when they cover up harms their AI caused, and is against broad licensing schemes that hand the market to the giants already ahead. Here the story shows a genuine discovery made through AI searching, with human scientists confirming it, and the company clearly separating the two roles. The division of labor between Claude’s search work and the human wet lab work is the whole point of the announcement.
Anthropic is positioning AI as a partner in discovery, not a replacement for it, and the company’s own statement that all lab work is performed by human scientists supports that framing. The paper wants to see startups able to use AI tools without being locked out by licensing schemes that only giants can afford. Readers should watch for any move toward broad federal regulation that would freeze today’s leaders in place.
Source material: “Anthropic says its biology lab has already found something big,” TechCrunch.
Get the Notebook.
The day's best stories and every fresh verdict, in plain English, in your inbox by seven. One email a day, no more.