Anthropic Reports Discovery of CRISPR-Like System, but Outside Scientists Say the Evidence Isn't In Yet

Anthropic has told the public that it discovered a biological system comparable to CRISPR, a claim that would mark a striking expansion for a company known for building large language models. But the experiments that would substantiate the finding are still pending, and no peer-reviewed results have appeared. The gap between the announcement and the evidence has drawn immediate pushback from researchers who work in the field.

The comparison to CRISPR carries enormous weight. CRISPR-Cas9, adapted from a bacterial immune mechanism, gave scientists a programmable way to cut and edit DNA, and it reshaped genetics, medicine and agriculture over the past decade. Any lab claiming a comparable system is effectively claiming a new foundational tool, not a marginal result. That is why outside experts expect data before interpretation, not the other way around.

One researcher summarized the situation bluntly, saying the experiments remain in the queue while the public relations effort is already live. That line captures the central complaint: the ordering of disclosure. In the life sciences, findings normally surface through preprints or journal papers, where methods, controls and replication details can be examined by people with relevant training. An announcement without those materials leaves reviewers with nothing to check.

There is also a structural question about who is making the claim. Anthropic's core expertise is AI research, not molecular biology, so a foray into laboratory science raises questions about the team behind the work, the institutions involved, and whether the company collaborated with established biology labs. None of those details have been laid out publicly in a form that specialists can assess. Until they are, the discovery remains a statement rather than a result.

For the broader AI industry, the episode tests how much credibility a company can borrow from its brand when it moves into a new scientific domain. AI firms have increasingly pointed to their models as tools for scientific discovery, and those claims often arrive with more press attention than verification. If Anthropic's finding holds up, it would strengthen that pitch considerably; if it does not, it becomes a cautionary example of announcing before proving.

What to watch is straightforward. The field needs the experiments finished, the methods released, and independent labs attempting replication of whatever mechanism Anthropic says it found. A preprint with raw data would answer more than any statement. Until that exists, the scientific community has a headline, a comparison to one of the most important tools in modern biology, and no way to confirm either.