Did Anthropic’s A.I. Really Make a Scientific Discovery on Its Own?

Live updates

1. 1d ago

   Anthropic Tests AI Scientific Autonomy in New Biology Lab

   Anthropic has expanded its operations from computer simulations to physical biological experiments through a new biology lab. The company aims to determine if AI agents can achieve genuine scientific discoveries and identify necessary capability limits. CEO Dario Amodei stated his goal is for Claude to eventually perform experiments by autonomously controlling lab equipment. This follows claims that Claude discovered a novel enzyme system called ARTs, a finding currently disputed by outside scientists who suggest the AI may have synthesized existing human research rather than generating original insights.

   Why it matters

   The debate centers on whether AI can move from pattern recognition to true scientific innovation. If AI can independently discover biological systems, it could accelerate drug discovery and genetic engineering. However, critics argue that current models may simply be repackaging known data.

   What is confirmed

   • Anthropic has established a biology lab to move beyond computer simulations into physical biological experiments.
   • Dario Amodei wants Claude to eventually perform experiments by autonomously controlling lab equipment.

   Still unconfirmed

   • Anthropic's AI synthesized prior human work rather than creating original scientific insights.

   What to watch next

   • Peer review of the ARTs enzyme discovery by independent biologists
   • Public disclosure of the specific constraints Anthropic is testing for AI capabilities
   • Evidence of Claude successfully controlling lab hardware without human intervention
   Sources used for this update (5)

   1. www.thedailystar.net — Why is Anthropic running a biology lab? | The Daily Star
   2. www.uniladtech.com — Anthropic says Claude has found something 'previously unknown ...
   3. jen.jiji.com — FlyDubai Flight, the co-pilot and the plan against Israel: "I wanted to avenge Gaza"
   4. jen.jiji.com — D'Alema: "Meloni lured the center-left into a trap"
   5. www.europesays.com — Did Anthropic’s A.I. Really Make a Scientific Discovery on ...
   confidence 80%
2. 3d ago

   Questions Arise Over Anthropic AI Discovery Claim

   Artificial intelligence company Anthropic faces scrutiny over claims that its Claude model independently discovered a novel enzyme system with CRISPR-like repeats called ARTs. While Anthropic announced the finding as a major breakthrough from its new biology lab, outside experts and scientists have questioned whether the artificial intelligence truly made the discovery on its own. Computational biologist Mario Rodriguez Mestre stated that he and his colleagues have studied these exact enzymes for four years and regularly used Anthropic models during their research, raising doubts about whether the system generated original insights or synthesized prior human work.

   Why it matters

   Technology companies frequently claim breakthroughs to demonstrate the expanding capabilities of artificial intelligence systems in specialized fields like biology. However, verifying independent discovery remains difficult when models have access to vast training data and interact directly with researchers writing code or drafting manuscripts. This incident highlights a growing debate in the scientific community regarding how to properly evaluate artificial intelligence autonomy versus human collaboration.

   What is confirmed

   • Anthropic announced that its biology lab and AI model Claude discovered a novel enzyme system with CRISPR-like repeats.
   • Computational biologist Mario Rodriguez Mestre stated that he and his colleagues had been studying these enzymes and associated molecules for four years.
   • Mario Rodriguez Mestre noted that he and his colleagues regularly used Anthropic artificial intelligence models to write code and draft manuscripts over the past three years.

   Still unconfirmed

   • Anthropic's AI model reasoned its way to the enzyme results entirely on its own without guidance from prior human research data.

   What to watch next

   • Full publication of Mario Rodriguez Mestre's research on the enzyme system
   • Further technical disclosures from Anthropic regarding the training data and prompting used by Claude
   Sources used for this update (18)

   1. Fox Business — Anthropic CEO says AI could cure most major diseases within next decade
   2. anthropic.com — Claude discovers a novel enzyme system with CRISPR-like repeats
   3. TechCrunch — Anthropic says its biology lab has already found something big
   4. The Verge — Anthropic’s biolab made a discovery it’s comparing to Crispr
   5. The New York Times — Did Anthropic’s A.I. Really Make a Scientific Discovery on Its Own?
   6. The Independent — Anthropic says Claude has found something potentially transformative in our bodies
   7. www.foxnews.com — Anthropic CEO predicts AI will cure most major diseases in a decade
   8. www.cnet.com — Did Anthropic Really Make a Big Biology Breakthrough? We Asked Scientists to Weigh In
   9. www.technologyreview.com — When can we say AI made a scientific discovery? | MIT Technology Review
   10. www.cnbc.com — OpenAI reveals Dots AI agents as protestors rally outside DevDay venue: Live updates
   11. jen.jiji.com — Real Madrid rejoins Efc, it's the end of the Super League project
   12. www.irishtimes.com — Did Anthropic’s artificial intelligence really make a scientific discovery on its own?
   confidence 90%