AI mirrors experimental science to uncover a mechanism of gene transfer crucial to bacterial evolution

Checked 2 Nov 2026. Open license (CC-BY, CC-BY-SA, CC0, public domain): full text indexed and used in synthesis.

Artificial intelligence (AI) models have been proposed for hypothesis generation, but testing their ability to drive high-impact research is challenging since an AI-generated hypothesis can take decades to validate. Here, we challenge the ability of a recently developed large language model (LLM)-based platform, AI co-scientist, to generate high-level hypotheses by posing a question that took years to resolve experimentally but remained unpublished: how could capsid-forming phage-inducible chromosomal islands (cf-PICIs) spread across bacterial species? Remarkably, the AI co-scientist's top-ranked hypothesis matched our experimentally confirmed mechanism: cf-PICIs hijack diverse phage tails to expand their host range. We critically assess its five highest-ranked hypotheses, showing that some opened new research avenues in our laboratories. We benchmark its performance against other LLMs and outline best practices for integrating AI into scientific discovery. Our findings suggest that AI can act not just as a tool but as a creative engine, accelerating discovery and reshaping how we generate and test scientific hypotheses.

• Penadés et al. 2025 give the hypothesis-generation claims of Gottweis et al. 2026's Co-Scientist a contamination-controlled test: they posed a question that took years to resolve experimentally but remained unpublished, and the system's top-ranked hypothesis matched the confirmed mechanism—cf-PICIs hijacking diverse phage tails—so the answer cannot have been retrieved from the published literature. Trace →
• Alampara et al. 2025 supply the reason contamination-controlled tests of AI scientists matter: model performance correlates strongly with online prominence, suggesting pattern matching rather than robust scientific understanding. Penadés et al. 2025's design—posing a question that took years to resolve experimentally but remained unpublished—is exactly the control that removes prominence-driven retrieval as an explanation of the co-scientist's success. Trace →
• AI-initiated biological discovery is now being validated against experimental ground truth that was never in the literature. Penadés et al. 2025 posed a question that took years to resolve experimentally but remained unpublished and found the AI co-scientist's top-ranked hypothesis matched the confirmed mechanism, while Yoon et al. 2026's autonomous Claude Code survey of 1.9 billion protein clusters surfaced a novel reverse-transcriptase family with expression evidence during phage infection. Trace →

• AI for science and agentic research