The mounting doubts over the independence of Claude’s research cast a shadow over the very first outputs of Anthropic’s biology lab, amid scientific controversy regarding the role of prior research in guiding artificial intelligence algorithms.
- Anthropic’s announcement and the ART system discovery
- Rodrigue Mistry’s testimony and details of prior collaboration
- Stance of leading scientists and specialized scientific journals
- Anthropic lab and life sciences research strategy
- Crucial questions regarding the future of autonomous discoveries
- FAQs
Anthropic’s announcement and the ART system discovery
Anthropic announced on September 23 that its artificial intelligence model, Claude, successfully identified and characterized a previously unknown enzymatic system within the DNA of bacterial phages, considering this achievement the first official research fruit of its new molecular biology lab. However, within days of the announcement, a sweeping wave of scientific doubts and questions erupted over whether this discovery was made entirely independently by the artificial intelligence, or whether the model relied on and was guided by data and detailed insights shared with it by external researchers from outside the company.
Anthropic explained that about 950 smart agents of Claude spent 21 continuous hours scanning and examining over 200,000 reverse transcriptase enzyme sequences in massive genetic databases, narrowing down thousands of candidate systems to a small group of promising threads. One of the agents successfully detected an unusual pattern of regularly spaced palindromic repeats next to the reverse transcriptase gene, in a structural layout resembling the famous CRISPR defense system used in gene editing. The company named this system “Array-Associated Reverse Transcriptases,” abbreviated as the “ART” system, consisting of the enzyme, an accessory gene of unknown function, and the repeat array. Despite the company acknowledging that the enzyme itself was previously known, it claimed that Claude was the first to recognize the broader landmarks and features of the system as a whole in a preliminary study that has not yet undergone peer review.
Rodrigue Mistry’s testimony and details of prior collaboration
In a surprising development that shook academic circles, The New York Times revealed on Sunday decisive statements made by computational biologist Dr. Rodrigue Mistry, in which he confirmed that he and his research team have been studying these enzymes and related molecules for a full four years. Dr. Mistry explained that his team had been using Anthropic’s artificial intelligence models for three years, and during that period exchanged and shared detailed insights, analyses, and precise scientific information about these enzymes with the company’s officials and researchers.
Mistry stated in candid remarks to the newspaper: “For me, the most important and crucial question is not whether the ART system was previously known; it was already known and identified in our research. My real concern centers on whether Claude deduced these findings truly independently and autonomously, or whether it was constrained and guided in its conclusions based on my own research.” Mistry added, expressing his concerns: “I deeply fear that this scientific data may have been used to train and enhance the capabilities of subsequent model versions without clear attribution to the foundational human efforts.”
Stance of leading scientists and specialized scientific journals
These doubts coincided with sober and conservative scientific evaluations from leading geneticists worldwide. Feng Zhang, a CRISPR pioneer at the Massachusetts Institute of Technology and the Broad Institute, described the research as “interesting and worthy of further study and investigation,” while the renowned scientific journal Nature stated that the true biological function of the ART system remains entirely unknown and mysterious, pointing to “very meager and sparse evidence” indicating that these arrays perform immune or regulatory functions similar to what the CRISPR system achieves.
Anthropic lab and life sciences research strategy
Anthropic established its life sciences research group in the spring of 2026 to test whether general language models could systematically and regularly accelerate biological discoveries. This facility, located in the San Francisco Bay Area, operates under low biosafety standards; it does not handle dangerous human pathogens at all, leaving the execution of physical and chemical experiments entirely to human scientists.
The company presents this lab as both a scientific endeavor and a practical experiment in artificial intelligence to study Claude’s behavior while conducting real scientific research. Eric Coder-Abrams, head of life sciences at Anthropic, told Nature that the goal lies in “institutionalizing and scaling up” exploratory research that has historically resulted in the creation of groundbreaking biological tools in modern medicine.
Crucial questions regarding the future of autonomous discoveries
The question remains hanging and open in scientific forums as to whether the ART system will prove its worth as a revolutionary gene editing tool in the future, and whether Claude arrived at it with complete independence and genuine computational curiosity, or whether it was actually retrieving the summary of scientific data and guidance instilled in it by human researchers who spent grueling years unravelling the secrets of these genetic molecules.
FAQs
Question: What is the scientific accusation or suspicion raised by researcher Rodrigue Mistry against Anthropic?
Answer: He noted that he shared his research regarding those enzymes with Anthropic, suggesting that Claude utilized that data rather than discovering it autonomously.
Question: What is the primary objective behind establishing Anthropic’s molecular biology lab?
Answer: To test general artificial intelligence’s ability to accelerate vital scientific discoveries and monitor its behavior in processing complex hypotheses.
Question: How did Nature magazine and independent scientists evaluate the effectiveness of the discovered ART system?
Answer: They confirmed that the system’s function remains unknown and there is insufficient evidence proving it possesses gene editing capabilities parallel to the CRISPR system.
