Story
September 25, 2026

Claude’s CRISPR-Like Find Sparks AI Hype—and a Reality Check

Anthropic says Claude helped uncover a previously uncharacterized enzyme system in bacteriophages after a rapid, agent-driven search. Scientists see an intriguing lead, but warn that its function, usefulness and significance remain unproven.

Anthropic formed its life-sciences research group and Bay Area wet lab in spring 2026 to test whether Claude could speed the painstaking work of finding uncharacterized biological systems. The company says the lab uses only BSL-1 and BSL-2 conditions, does not handle human-infecting pathogens, and leaves all physical experiments to human scientists.

Its first public result came after Claude agents searched a vast DNA database for unusual reverse transcriptases. Anthropic says roughly 950 agents used 210 million tokens over 21 hours, narrowing more than 200,000 sequences to candidates for human review. One agent flagged a repeating DNA pattern beside an odd-looking enzyme in bacteriophages, viruses that infect bacteria. Lab follow-up identified what Anthropic calls array-associated reverse transcriptases, or ARTs.

The company’s case is deliberately ambitious but not absolute: ART combines features seen in a small number of programmable systems that can cut, copy or paste DNA. Yet Anthropic concedes it does not know ART’s primary function. CEO Dario Amodei called it a suspected new gene-editing mechanism, while stressing that its “precise function, biotechnological utility (if any), or level of significance is not yet clear.”

That caveat is central to the scientific pushback. Philip Kranzusch, a microbiologist at Harvard Medical School, said the result was “not yet a breakthrough” but “a wrinkle on what we know in the field.” The technical work has not yet been submitted to a scientific journal, leaving broader validation still ahead.

Anthropic says it shared the result early both to show what Claude can do and to give researchers visibility into its work. Critics see a more complicated backdrop: the company is trying to recruit scientists and expand into drug discovery as AI firms increasingly use research results to demonstrate their models’ value. The CRISPR comparison may be a useful signal of what ART could become—but, for now, it is not proof that it will become anything at all.

Story coverage