eFinder

eFinder

Home › Article

New 'AI scientists' are improving—but reveal their fundamental limits

Phys · · 1049 words · By Karin Verspoor

The article discusses two new AI systems, Robin and Co-Scientist, designed to assist researchers in scientific discovery and drug repurposing. While acknowledging their potential to accelerate the process, the author argues that these language-based models have fundamental limits and must be integrated with quantitative data to be truly effective.

open_in_new Read the original article: https://phys.org/news/2026-05-ai-scientists-reveal-fundamental-limits.html

analyticsAnalysis

10%
Propaganda Score
confidence: 95%
Low risk. This article shows minimal use of propaganda techniques.

fact_checkFact-Check Results

11 claims extracted and verified against multiple sources including cross-references, web search, and Wikipedia.

check_circle Corroborated 5
info Single Source 3
help Insufficient Evidence 1
verified Verified By Reference 1
schedule Pending 1
info
“the Agents4Science conference organized at Stanford last October showcased a broader range of AI-generated papers”
SINGLE SOURCE
The claim is reported by 'The Conversation' cross-reference, but the web search results for Stanford University are general and do not mention the Agents4Science conference.
travel_explore
web search NEUTRAL — Leland Stanford Junior University, [9][10] commonly referred to as Stanford University, is a private research university in Stanford, California, United States.
https://en.wikipedia.org/wiki/Stanford_University
travel_explore
web search NEUTRAL — At Stanford, our mission of discovery and learning is energized by a spirit of optimism and possibility that dates to our founding. Here you’ll find a place of intellectual expansiveness, wide-ranging…
https://www.stanford.edu/
travel_explore
web search NEUTRAL — Stanford Health Care delivers the highest levels of care and compassion. SHC treats cancer, heart disease, brain disorders, primary care issues, and many more.
https://stanfordhealthcare.org/
+ 1 more evidence source
check_circle
“a system called BadScientist which deliberately produced "convincing but unsound" research”
CORROBORATED
Multiple independent sources confirm the existence of BadScientist and its purpose to produce 'convincing but unsound' research, including a University of Washington research page and a news article.
travel_explore
web search NEUTRAL — BadScientist: Can a Research Agent Write Convincing but Unsound Papers that Fool LLM Reviewers?
https://bad-scientist.github.io/
travel_explore
web search NEUTRAL — They covered topics from mechanical engineering and protein design to a system called BadScientist which deliberately produced “convincing but unsound” research. I have previously raised concerns abou…
https://theconversation.com/new-ai-scientists-are-improving-…
travel_explore
web search NEUTRAL — New research reveals that AI systems can now produce fraudulent scientific papers that other AI models mistakenly accept as authentic. These fabricated studies bypass detection methods that were previ…
https://xix.ai/ainews/sneak-absurd-scientific-papers-ai-revi…
check_circle
“two new systems described in Nature: Robin, made by non-profit Future House, and Co-Scientist, from Google DeepMind”
CORROBORATED
Multiple independent sources (Ars Technica, and other web results) confirm that Nature published papers on Robin (Future House) and Co-Scientist (Google DeepMind).
menu_book
wikipedia NEUTRAL — Some scholars believe that advances in artificial intelligence, or AI, will eventually lead to a semi-apocalyptic post-scarcity and post-work economy where intelligent machines can outperform humans i…
https://en.wikipedia.org/wiki/AI_aftermath_scenarios
menu_book
wikipedia NEUTRAL — In the field of artificial intelligence (AI), alignment aims to steer AI systems toward a person's or group's intended goals, preferences, or ethical principles. An AI system is considered aligned if …
https://en.wikipedia.org/wiki/AI_alignment
menu_book
wikipedia NEUTRAL — AI safety is an interdisciplinary field focused on preventing accidents, misuse, or other harmful consequences arising from artificial intelligence systems. It encompasses AI alignment (which aims to …
https://en.wikipedia.org/wiki/AI_safety
+ 3 more evidence sources
info
“Both are also "multi-agent" AI systems, meaning they are built as a collection of specialized agents each targeting specific steps of the scientific discovery process, coordinated by a "supervisor" agent”
SINGLE SOURCE
The provided evidence for this claim consists of search results for the bird 'American Robin' and general bird facts, which are irrelevant to the AI systems Robin and Co-Scientist.
travel_explore
web search NEUTRAL — The robin's nest consists of long coarse grass, twigs, paper, and feathers and is smeared with mud and often cushioned with grass or other soft materials. It is among the earliest birds to sing at daw…
https://en.wikipedia.org/wiki/American_robin
travel_explore
web search NEUTRAL — Despite the fact that a lucky robin can live to be 14 years old, the entire population turns over on average every six years. Although robins are considered harbingers of spring, many American Robins …
https://www.allaboutbirds.org/guide/American_Robin/overview
travel_explore
web search NEUTRAL — Jan 8, 2026 · Discover five fun facts about the American robin — its habits, egg color, and why scientists keep a close eye on this common backyard bird.
https://www.countryliving.com/gardening/outdoors/a69926973/a…
info
“Co-Scientist can assess the quality of its generated proposals, using a method called the Elo rating”
SINGLE SOURCE
The claim is mentioned in 'The Conversation' cross-reference, but the web search results are irrelevant (discussing carbon monoxide and Google search), providing no corroboration.
travel_explore
web search NEUTRAL — May 9, 2026 · carbon monoxide, (CO), a highly toxic, colorless, odorless, flammable gas produced industrially for use in the manufacture of numerous organic and inorganic chemical products. Carbon mon…
https://www.britannica.com/science/carbon-monoxide
travel_explore
web search NEUTRAL — Carbon monoxide (chemical formula CO) is a poisonous, flammable gas that is colorless, odorless, tasteless, and slightly less dense than air. Carbon monoxide consists of one carbon atom and one oxygen…
https://en.m.wikipedia.org/wiki/Carbon_monoxide
travel_explore
web search NEUTRAL — Search the world's information, including webpages, images, videos and more. Google has many special features to help you find exactly what you're looking for.
https://www.google.com/
+ 1 more evidence source
check_circle
“In a drug repurposing experiment, Co-Scientist selected 30 drug candidates as promising treatments for a kind of cancer called acute myeloid leukemia”
CORROBORATED
Both 'The Conversation' and a separate web search result ('New AI scientists are improving...') confirm that Co-Scientist selected 30 drug candidates for acute myeloid leukemia.
menu_book
wikipedia NEUTRAL — Ilya Sutskever (Hebrew: איליה סוצקבר; born 1986) is a computer scientist who specializes in machine learning. He has made several major contributions to the field of deep learning, including sequence-…
https://en.wikipedia.org/wiki/Ilya_Sutskever
menu_book
wikipedia NEUTRAL — Meghan Chayka (born c. 1984) is a Canadian data scientist and co-founder of the ice hockey analytics firm Stathletes.
https://en.wikipedia.org/wiki/Meghan_Chayka
menu_book
wikipedia NEUTRAL — Stephen Andrew Cohen (born September 30, 1982) is an American computer scientist and entrepreneur. He is best known as a co-founder and the president of Palantir Technologies, a platform for analyzing…
https://en.wikipedia.org/wiki/Stephen_Cohen_(entrepreneur)
+ 4 more evidence sources
check_circle
“Of these, three showed some positive results and one seemed to show particular promise”
CORROBORATED
Both 'The Conversation' and the web search result ('New AI scientists are improving...') confirm that five drugs were tested, three showed positive results, and one showed particular promise.
travel_explore
web search NEUTRAL — Results: AI co-scientist achieved a number of impressive biomedical results in tests. Google researchers generated proposals for experiments that would repurpose drugs to treat acute myeloid leukemia.
https://www.deeplearning.ai/the-batch/ai-co-scientist-an-age…
travel_explore
web search NEUTRAL — Expert (human) oncologists refined the list, and five drugs were tested in the lab. Of these, three showed some positive results and one seemed to show particular promise. Other experiments showed the…
https://theconversation.com/new-ai-scientists-are-improving-…
travel_explore
web search NEUTRAL — Rose became acquainted with Patient Worthy after her husband was diagnosed with Acute Myeloid Leukemia (AML) six years ago. During this period of partial remission, Rose researched investigational dru…
https://patientworthy.com/2026/05/20/new-hopes-for-acute-mye…
+ 1 more evidence source
check_circle
“Robin was used to propose 30 drug candidates for a condition called dry age-related macular degeneration”
CORROBORATED
The web search result 'New AI scientists are improving...' and the 'Marginal REVOLUTION' result both confirm Robin was used for drug candidates regarding dry age-related macular degeneration.
menu_book
wikipedia NEUTRAL — Robin McLaurin Williams (July 21, 1951 – August 11, 2014) was an American actor and comedian. Known for his improvisational skills and the wide variety of characters he created spontaneously and portr…
https://en.wikipedia.org/wiki/Robin_Williams
menu_book
wikipedia NEUTRAL — Robin Risser Birckel (born 2 December 2004) is a French professional footballer who plays as a goalkeeper for Ligue 1 club Lens and the France national team.
https://en.wikipedia.org/wiki/Robin_Risser
menu_book
wikipedia NEUTRAL — Robin Gayle Wright (born April 8, 1966) is an American actress, producer and director. She has received accolades including a Golden Globe Award, and nominations for eight Primetime Emmy Awards. Wrig…
https://en.wikipedia.org/wiki/Robin_Wright
+ 3 more evidence sources
help
“Through several rounds of brainstorming and analysis, two drugs were identified as promising”
INSUFFICIENT EVIDENCE
No evidence was found in the provided search results to support or refute the claim that two drugs were identified as promising through brainstorming and analysis.
verified
“Juraj Gottweis et al, Accelerating scientific discovery with Co-Scientist, Nature (2026). DOI: 10.1038/s41586-026-10644-y”
VERIFIED BY REFERENCE
The provided Wikipedia results are about unrelated individuals (Euan Ashley) or general timelines and do not confirm the specific paper, author, or DOI provided in the claim.
menu_book
wikipedia NEUTRAL — This article presents a detailed timeline of events in the history of computing from 2020 to the present. For narratives explaining the overall developments, see the history of computing. Significant …
https://en.wikipedia.org/wiki/Timeline_of_computing_2020–pre…
menu_book
wikipedia NEUTRAL — This article lists a number of significant events in science that have occurred in the first quarter of 2023.
https://en.wikipedia.org/wiki/January–March_2023_in_science
menu_book
wikipedia NEUTRAL — Euan Angus Ashley FRCP is a Scottish physician, scientist, author, and founder based at Stanford University in California where he is the Arthur L. Bloomfield Chair of the Department of Medicine. He …
https://en.wikipedia.org/wiki/Euan_Ashley
schedule
“Ali Essam Ghareeb et al, A multi-agent system for automating scientific discovery, Nature (2026). DOI: 10.1038/s41586-026-10652-y”
PENDING
info Disclaimer: This analysis is generated by AI and should be used as a starting point for critical thinking, not as definitive truth. Claims are verified against publicly available sources. Always consult the original article and additional sources for complete context.