Anthropic has published the first results from the life-sciences lab it opened in the Bay Area in spring 2026, and the claim it is making is a large one: that its Claude agents autonomously identified a previously uncharacterised enzyme system whose architecture echoes CRISPR.

According to the company, human involvement was limited to the initial prompt and the laboratory work. About 950 Claude agents then spent 21 hours and 210 million tokens combing a large DNA database for unusual reverse transcriptases (RTs) — enzymes that copy RNA back into DNA. The agents gathered more than 200,000 RTs, narrowed them to 3,500 candidate systems and produced detailed, human-readable reports on the 20 most compelling. One agent, reading raw sequence data, spotted a tandem repeat array sitting next to an odd-looking RT and wrote in its report: "…I can see by eye a tandem repeat array … that's a CRISPR-like … repeat array?!"

Anthropic calls the system ART — array-associated reverse transcriptases. It turns up mainly in bacteriophages, the viruses that infect bacteria, and has three parts: the RT itself, a partner gene of unknown function and a long array of evenly spaced DNA repeats. The RT had been seen in earlier studies; Anthropic says Claude appears to be the first to notice the repeat array and the accessory protein that define the system. Early experiments show the array is transcribed into a set of short RNAs, hinting at CRISPR-like programmability. What ART actually does is still unknown, and the company has released only a preprint.

Reaction has split along familiar lines. Feng Zhang, a CRISPR pioneer at MIT and the Broad Institute, called it "an exciting example of how AI agents can contribute to biological discovery" and said the RNA-repeat arrays associated with reverse transcriptases are "genuinely intriguing" and merit further investigation. Lucas Harrington, who earned his PhD under Nobel laureate Jennifer Doudna and co-founded Mammoth Biosciences — a company that itself uses genome mining to hunt for CRISPR systems — was blunt: the method has been around for decades and similar systems have been known since 2008. The hard part, he argued, is working out what a system does, and Anthropic has not shown that; presenting early results as a major discovery is not helpful.

That gap is the point. Anthropic is explicit that the function of ART is unresolved and that further experiments are under way. The announcement lands as the company tries to recruit biologists and expand into drug discovery, alongside CEO Dario Amodei's public claim that AI could cure most major diseases within five to ten years. For now, the proof of the method is a flagged anomaly and a preprint — not a tool anyone can use.

Sources