aiqbench analysis
This controlled study reveals that mid-2025 LLMs struggle to replicate expert-level literature searches in physics, showing less than 6% overlap with human selections due to significant metadata hallucinations. While current models require systematic verification for reference reliability, the analysis suggests a promising trajectory where future iterations could effectively complement rather than replace researchers. The findings underscore that AI's utility in scientific discovery currently lies in augmentation and speed-up of tasks like initial screening, not autonomous expert-level curation.
AI's Capability in Assisting Scientific Research in Physics, Astrophysics, and Cosmology I: Literature Review
arXiv:2607.25672v1 Announce Type: cross Abstract: We investigate how well large language models (LLMs) can assist with literature reviews for scientific research. We perform a controlled study of eight expert-conceived research projects across the areas of physics, astrophysics, and cosmology. Each project has a defined background and goal, and human experts and AI prompters are asked to perform identical literature review tasks in parallel. We compare the relevant literature selected by humans with that selected by mid-2025 LLMs (ChatGPT-4o, ChatGPT Deep Research, and Gemini). We find the overlap between human- and AI-selected references to be small ($<$6\%), indicating that AI models do not yet reproduce a competent expert search on their own, though they have the potential to complement literature searches by humans. We then assess the reliability and completeness of AI-generated candidate references, distinguishing two types of hallucination: fabrications (references to nonexistent papers) and metadata mismatches (real papers with one or more incorrect fields). We find that while fabricated references make up 3\% of the AI-generated references, 64\% are real papers with at least one incorrect field (title, author, year, journal, DOI, or link), indicating that the mid-2025 models require systematic verification. However, the performance is significantly improved for the 2026 model ChatGPT Pro 5.5, with a single-project test showing zero fabrication or metadata mismatches.
Authors:Anamaria Hell, Kateryna Vovk, Veena Krishnaraj, Jia Liu, Kosuke Aizawa, Adrian E. Bayer, Linda Blot, Jessica Cowell, Suyog Garg, Jonathan Grée, Ben Horowitz, Masaya Ichikawa, Kanyuni Iemoto, Keigo Kondo, Zacharie Lorsin, Kevin McCarthy, Jamie Robinson, Miguel Ruiz-Granda, Leander Thiele, Ievgen Vovk, Mingshen Zhou
Abstract:We investigate how well large language models (LLMs) can assist with literature reviews for scientific research. We perform a controlled study of eight expert-conceived research projects across the areas of physics, astrophysics, and cosmology. Each project has a defined background and goal, and human experts and AI prompters are asked to perform identical literature review tasks in parallel. We compare the relevant literature selected by humans with that selected by mid-2025 LLMs (ChatGPT-4o, ChatGPT Deep Research, and Gemini). We find the overlap between human- and AI-selected references to be small ($<$6\%), indicating that AI models do not yet reproduce a competent expert search on their own, though they have the potential to complement literature searches by humans. We then assess the reliability and completeness of AI-generated candidate references, distinguishing two types of hallucination: fabrications (references to nonexistent papers) and metadata mismatches (real papers with one or more incorrect fields). We find that while fabricated references make up 3\% of the AI-generated references, 64\% are real papers with at least one incorrect field (title, author, year, journal, DOI, or link), indicating that the mid-2025 models require systematic verification. However, the performance is significantly improved for the 2026 model ChatGPT Pro 5.5, with a single-project test showing zero fabrication or metadata mismatches.
Instrumentation and Methods for Astrophysics (astro-ph.IM); Cosmology and Nongalactic Astrophysics (astro-ph.CO); Computation and Language (cs.CL); General Relativity and Quantum Cosmology (gr-qc)
arXiv:2607.25672 [astro-ph.IM]
(or arXiv:2607.25672v1 [astro-ph.IM] for this version)
https://doi.org/10.48550/arXiv.2607.25672 arXiv-issued DOI via DataCite (pending registration)
Full text under the publisher's open licence · 289 words
Related coverage
AI's Capability in Assisting Scientific Research in Physics, Astrophysics, and Cosmology II: Project Planning and Proposal Evaluation
arXiv · Jul 29, 2026
A New Era of Discovery: Google Research at I/O 2026
Google Research · May 28, 2026
[AINews] not much happened today
Latent Space · Sep 18, 2026
XtalPi Holdings Limited: XtalPi Holdings Announces 2026 Interim Results
finanznachrichten.de · Aug 19, 2026