Anthropic Says Claude Discovers New CRISPR-Like Enzyme System In Bacterial DNA

Anthropic says its Claude AI model has helped scientists discover a previously uncharacterised enzyme system in bacterial DNA that shows some similarities to CRISPR. The system, called array-associated reverse transcriptases (ARTs), was identified after Claude analysed more than 200,000 DNA sequences over about 21 hours. Anthropic says the system's biological function is still unknown and cautions that it has not been established as a new gene-editing technology. Scientists have welcomed the potential of AI-assisted biological research while warning against prematurely calling the discovery a “next CRISPR.”

CLAUDE IDENTIFIES PREVIOUSLY UNCHARACTERISED BIOLOGICAL SYSTEM

Artificial intelligence company Anthropic says its Claude AI model has helped identify a previously uncharacterised enzyme system in bacterial DNA with features that resemble the biological patterns associated with CRISPR.
Anthropic announced the discovery on Wednesday as part of the launch of its new life sciences research group and laboratory in the San Francisco Bay Area.
The company said Claude identified the system after being given a broad research task to search a massive database of DNA sequences for unusual reverse transcriptases.
Anthropic said the AI agents worked for about 21 hours, analysing more than 200,000 reverse transcriptases and narrowing thousands of candidates down to a small number for detailed investigation.

THE SYSTEM HAS BEEN NAMED ART

Anthropic researchers have named the newly identified system array-associated reverse transcriptases, or ARTs.
The system is mainly found in bacteriophages, viruses that infect bacteria.
It consists of a reverse transcriptase, a neighbouring partner gene and a long array of evenly spaced DNA repeat sequences.
Anthropic said the arrangement of these repeats resembles the arrays found in CRISPR systems.
However, the company stressed that researchers do not yet know what ART actually does.
Laboratory experiments are continuing to determine its biological function.

CLAUDE SPOTTED A PATTERN RESEARCHERS HAD NOT IDENTIFIED

Anthropic said the underlying reverse transcriptase had previously been identified by scientists.
What Claude apparently noticed was the unusual combination of the enzyme with a neighbouring gene and a repeated DNA sequence pattern.
The AI then examined the DNA in greater detail, counted and compared the repeats, searched existing scientific literature and generated a report for human researchers.
Scientists subsequently investigated the candidate in the laboratory and confirmed that it represented a previously uncharacterised biological system.
Anthropic said this illustrates how AI could help researchers examine enormous biological datasets and identify patterns that might otherwise take humans weeks or months to find.

WHY THE CRISPR
COMPARISON MATTERS
CRISPR is one of the most

important technologies to emerge from modern molecular biology.
The naturally occurring CRISPR system is part of the defence mechanisms used by bacteria against invading viruses. Scientists later learned how to adapt parts of the system into powerful gene-editing tools.
The technology allows researchers to make targeted changes to DNA and has contributed to the development of treatments for certain genetic diseases.
Scientists Emmanuelle Charpentier and Jennifer Doudna received the 2020 Nobel Prize in Chemistry for developing a method of genome editing based on CRISPR-Cas9.
Because of CRISPR's importance, the discovery of another biological system with some similar characteristics has attracted considerable attention.

ANTHROPIC SAYS IT IS NOT YET A ‘NEW CRISPR’

Anthropic has been careful not to claim that ART is a replacement for CRISPR.
The company says the system has characteristics found together in only a small number of other programmable biological systems, but its function remains unknown.
That distinction is important because a naturally occurring CRISPR-like sequence is not automatically a usable gene-editing technology.
Researchers would first have to understand how ART works, determine whether it can be controlled, and establish whether it can perform useful biological functions.

SCIENTISTS SEE PROMISE BUT URGE CAUTION

Feng Zhang, one of the pioneers of CRISPR genome editing, described the finding as an exciting example of AI contributing to biological discovery.
Stanley Qi, a Stanford University bioengineering professor, also described the reported discovery as exciting, particularly because of Claude's ability to recognise an unusual biological pattern and investigate it systematically.
However, not all scientists are convinced that the discovery represents a potential breakthrough in gene editing.
Microbiologist Kevin Blake of Washington University School of Medicine cautioned against describing the system as the “next CRISPR”.
He noted that nature contains enormous numbers of biological systems that remain undiscovered and that a CRISPR-like sequence does not necessarily mean a system can be developed into a medical treatment.

AI IS CHANGING HOW SCIENTISTS SEARCH DNA

The discovery highlights one of the areas where AI could have a major impact on biological research.
DNA databases contain enormous amounts of information, much of which has not been fully understood.
Traditional genome mining requires scientists to search these databases, identify unusual sequences, compare them with existing knowledge and develop hypotheses about what they might do.
Anthropic says Claude can accelerate that process by examining large numbers of candidates simultaneously and producing reports that human scientists can then evaluate.
In the ART project, Anthropic said Claude agents examined more than 200,000 reverse transcriptases and identified thousands of potential systems before narrowing them down for human review.

HUMANS STILL CARRIED OUT THE LABORATORY WORK

Despite describing the discovery as autonomous, Anthropic said human researchers remained responsible for the experimental work.
The AI searched and analysed DNA sequences, generated hypotheses and helped interpret information.
Scientists then selected promising candidates and tested them in the laboratory.
Anthropic said its life sciences laboratory works at low biosafety levels and does not handle pathogens capable of infecting humans.
The distinction is important because AI-generated biological hypotheses still need to be tested and verified experimentally before scientists can establish whether they are correct.

ANTHROPIC EXPANDS INTO BIOLOGICAL RESEARCH

The discovery comes as Anthropic expands its use of Claude beyond conventional software and knowledge work.
The company established a dedicated life sciences research group in 2026 to investigate how AI can help scientists identify proteins, understand DNA and accelerate biological discovery.
Anthropic has also introduced specialised AI programmes for life sciences organisations, while maintaining stronger safeguards around potentially dangerous biological research.
The company says its broader goal is to develop a research model in which AI agents and human scientists work together throughout the discovery process.

AI'S GROWING ROLE IN MEDICAL RESEARCH

Anthropic CEO Dario Amodei has previously argued that increasingly capable AI systems could accelerate medical research and help scientists develop treatments for serious diseases.
The ART discovery provides an example of the type of early-stage scientific work that AI companies believe their systems could increasingly perform.
But the discovery is still far from becoming a medical technology.
Researchers must first establish what ART does, determine whether its molecular machinery can be manipulated and assess whether it has any practical applications.

CRISPR HAS ALREADY CHANGED MEDICINE

The excitement surrounding the discovery partly reflects how dramatically CRISPR has changed medicine.
Researchers have developed gene-editing treatments for certain inherited diseases, including sickle cell disease.
In 2025, the Children's Hospital of Philadelphia also reported treating an infant with a personalised CRISPR-based therapy for a rare metabolic disorder.
These developments demonstrate why the discovery of new biological systems capable of interacting with DNA could eventually be important.
But any possible medical application of ART remains speculative at this stage.

DISCOVERY DOES NOT YET MEAN A NEW GENE-EDITING TOOL

For now, the most significant aspect of the announcement may be the demonstration that an AI system can search huge biological datasets, recognise unusual patterns and generate hypotheses that scientists consider worth investigating.
Whether ART becomes a practical gene-editing technology is unknown.
Anthropic's researchers are continuing experiments to understand the system's primary function.
The company says it is sharing the finding early so that other scientists can examine the discovery and potentially contribute to understanding how the newly identified system works.

A NEW ERA OF AI-ASSISTED DISCOVERY

The ART discovery adds to a growing body of work exploring whether AI can move beyond helping scientists analyse existing knowledge and instead contribute to finding previously unknown biological systems.
If such systems can reliably identify meaningful patterns across enormous datasets, they could accelerate research in genetics, medicine, chemistry and other fields.
For now, however, ART remains a scientific mystery rather than a proven successor to CRISPR.
The next stage will depend on laboratory experiments showing exactly what the system does and whether its unusual molecular structure can eventually be turned into a useful scientific or medical tool.