AI ModelsSeptember 24, 2026

Anthropic says Claude found a new enzyme system

It has CRISPR-like repeats, but what it does is still unknown.

Anthropic has set up a life sciences research group with its own laboratory. On September 23, it said the group’s first result came from Claude: an overlooked enzyme system in bacteriophages, the viruses that infect bacteria. Anthropic calls it ART and posted a preprint on its own website.

By Deep-Tech ResearcherBIOLOGY & FOUNDATION MODELS

Published: September 24, 2026, 18:05 CEST
Reading time: 5 min
Evidence: 6 sources · 1 primary

This story was generated by AI systems under the editorial rules of Silicon AI News and checked against the sources it links. Found a mistake? Email editor@siliconainews.com. How we report

In brief

Anthropic said on September 23 that Claude agents, searching huge DNA databases, spotted an overlooked enzyme system in viruses that infect bacteria. It has evenly spaced repeats like those in CRISPR. Anthropic’s own lab made a first test, but the work is not peer-reviewed.

New to this? Read it in simple words

Anthropic says its AI, Claude, found an overlooked group of genes in viruses that infect bacteria.
The genes sit next to repeated pieces of DNA, a pattern that looks like CRISPR.
Anthropic’s own lab did a first test. No outside scientists have checked the work yet.
Nobody knows yet what the system does, or if it will ever be useful.

Words to know

CRISPR: A natural defence system in bacteria that scientists turned into a tool for editing genes.
Enzyme: A protein that speeds up a chemical reaction in a living thing.
Peer review: When independent experts check a study before a journal publishes it.

What Claude found

ART stands for array-associated reverse transcriptases. It has three parts: an enzyme that copies RNA into DNA, a partner gene with an unknown role, and a row of evenly spaced DNA repeats.

Those spaced repeats look like the arrays in CRISPR, the bacterial defence system behind today’s best-known gene-editing tools. The resemblance is in the layout. Anthropic has not shown that ART cuts or edits DNA.

The enzyme itself was already known from earlier studies of a large phage. Anthropic says Claude was the first to notice the repeat array and the partner protein next to it.

Chief executive Dario Amodei wrote on X that a Stanford team had earlier found a system that is similar in some ways.

Sources1 3 4

01AGENT SEARCH~950 AGENTS · 21 HOURS

02CANDIDATES3,500 → 20 REPORTS

03FIRST LAB TESTSHORT RNAs · OWN LAB

04PEER REVIEWNOT YET

LAB CLAIM 02

Claude spotted the pattern. What the system does is still unknown.

All lab results so far come from Anthropic’s own lab. The preprint is not peer-reviewed.

How the search worked

Anthropic says about 950 Claude agents searched for 21 hours and used around 210 million tokens. They screened more than 200,000 enzymes, narrowed them to 3,500 candidates, and wrote 20 reports for scientists.

According to the preprint, as reported by The Next Web, the agents ran on Claude Mythos 5. When Anthropic repeated the same search ten more times, every rerun missed the repeat array.

People wrote the first instructions, reviewed the reports, chose what to test, and did all the lab work. Claude searched the data, picked candidates, wrote the reports, and helped interpret the results.

Anthropic’s own scientists ran the first test in its Bay Area lab. They found that the repeat array is turned into distinct short RNAs. No outside lab has tested the system yet.

Sources1 4 6

Not a CRISPR rival yet

The work has not been peer-reviewed. The preprint is posted only on Anthropic’s website, and no data or code has been released.

What ART does in nature is unknown. So is whether its enzyme is active, and whether it could ever become a useful tool.

Stanford bioengineer Stanley Qi told Al Jazeera the finding is very exciting. Microbiologist Kevin Blake of Washington University School of Medicine was more careful: “There’s nothing to indicate this is a rival to CRISPR-the-technology”.

Some early reports described ART as a new gene-editing tool. Amodei wrote only that Anthropic suspects this, and that the system’s function and importance are not yet clear.

Anthropic says its lab works only at biosafety levels 1 and 2 and handles no human pathogens.

Sources1 2 3 5

Sources

Every fact in this story comes from the sources below. Open them to check our work.

1
Primary source · September 23, 2026Claude discovers a novel enzyme system with CRISPR-like repeats Anthropic
2
Research · September 23, 2026Anthropic says Claude AI helped discover novel enzyme system Reuters via KSL
3
Research · September 23, 2026Anthropic says its biology lab has already found something big TechCrunch
4
Research · September 23, 2026Anthropic says Claude found a new enzyme system with CRISPR-like repeats The Next Web
5
Research · September 24, 2026AI model Claude discovers CRISPR-like enzyme system, Anthropic says Al Jazeera
6
Research · September 18, 2026Anthropic is operating a lab that conducts biology experiments TechCrunch

How we checked this story

We read Anthropic’s announcement, then compared Reuters, TechCrunch, The Next Web, and Al Jazeera. We could not open the full preprint, so details from it come from The Next Web’s reporting. All lab results are Anthropic’s own, and no outside lab has repeated them.

Story mapLAB CLAIM

How Claude ART Discovery connects

CRISPR-like Systems
Claude ART Discovery found repeats that look like CRISPR arrays , connected to CRISPR-like Systems
Research Acceleration
Claude ART Discovery about 950 agents searched in 21 hours , connected to Research Acceleration
Also in 1 other story
Biomedical Validation
Claude ART Discovery no outside lab test or peer review yet , connected to Biomedical Validation
Also in 1 other story
OpenAI Maths Claim
Claude ART Discovery another AI discovery claim awaiting outside checks , connected to OpenAI Maths Claim
Also in 1 other story

The story so far

Earlier stories on the same thread, oldest first.

AI ModelsSep 22

Claude Opus 5.5 cuts token prices by 20%

It tops one independent index. Most other claims come from Anthropic.

Anthropic released Claude Opus 5.5 on September 22 at $4 per million input tokens and $20 per million output tokens. It says the model matches its top Fable 5.1 on most work, and one independent index ranks it first.

5 min read

AI ModelsSep 23

OpenAI says a new model solved over 100 open maths problems

It has not named or published any of them yet.

In a September 21 post, OpenAI said an internal model in training since August 28 has solved more than 100 long-standing problems. It named none and published no proofs. It also set up an unpaid advisory group of nine leading mathematicians, hosted at the Institute for Advanced Study.

5 min read

Latest news

All news

Policy & LawOct 1

California bars firing or disciplining workers by AI alone

Newsom signed the No Robo Bosses Act with a dozen other bills on September 30. From July 2027, a human must check any discipline or firing that leans mainly on an automated system, and workers must be told in writing.

Governor Gavin Newsom signed SB 947, the “No Robo Bosses Act”, on September 30, along with 12 other bills his office grouped as AI and worker protections. From July 1, 2027, California employers may not rely solely on an automated decision system to discipline or fire workers. When such a system was the main basis, a human must corroborate the decision, and the worker must get a written notice.

4 min read

Safety & SecurityOct 1

OpenAI blames Moonshot-linked users for a copying campaign

OpenAI says the accounts tried to pull its models’ hidden reasoning in July, until it shut the campaign down. It published no evidence for the link to the Kimi maker, which had not responded publicly.

OpenAI said on September 30 that it had stopped a coordinated campaign to extract the hidden reasoning of its models. It attributes a core cluster of the activity to people associated with Moonshot AI, the Chinese developer of Kimi. The campaign peaked on July 24 and 25 with 16,000 attempts from more than 4,000 users, and OpenAI says it was fully disrupted by July 28.

3 min read

Safety & SecurityOct 1

DeepMind watermarks AI-designed proteins for biosecurity

SynthID Bio hides a signature in a protein’s building blocks, and lab tests found marked designs worked as well as unmarked ones. It could help DNA makers tell designs from trusted AI tools apart from the rest.

Google DeepMind introduced SynthID Bio on September 30, a way to watermark proteins designed by AI without breaking them. In lab tests on three targets, watermarked protein binders matched unwatermarked ones on how often they worked, how tightly they bound and how varied they were, DeepMind says. It is sharing the code, lab data and model weights with researchers.

4 min read

Policy & LawOct 1

Connecticut’s AI rules on layoffs and whistleblowers start

From October 1, employers must say whether AI played a part in mass layoffs, and makers of the biggest AI models may not punish staff who warn of catastrophic risks. AI subscriptions and AI-made media get new rules too.

Large parts of Connecticut’s new AI law took effect on October 1. Employers that file mass-layoff notices must now tell the state whether the layoffs relate to their use of AI. Developers of the biggest AI models may not retaliate against staff who report catastrophic risks, and AI subscription services must spell out their limits before customers pay.

4 min read

BusinessOct 1

ElevenLabs is now valued at $22 billion after a share sale

Employees sold $300 million of stock at twice February’s valuation, and the company raised no new money. ElevenLabs says businesses now bring in 55% of its revenue, and voice agents are its fastest-growing product.

ElevenLabs, the AI voice company, said on September 30 that a $300 million share sale by employees values it at $22 billion. Wellington and T. Rowe Price led the deal, and investors including EQT, Goldman Sachs and Singapore’s GIC bought in for the first time. The valuation is double the $11 billion of its funding round in February.

3 min read