Kathryn Hulick · Oct 21, 2026
Science News: Paper2Agent turns scientific papers into reusable AI tools
Markdown · JSON · Original source
reports that Paper2Agent packages a scientific paper’s code, methods and data into tools an AI assistant can use with new data. The system also tests the tools against reference outputs and provides instructions for using them. In tests of 100 computational biology papers, researchers converted 74 into working tool sets. For converted papers, the system ran analyses and answered questions about results more accurately and more than twice as fast as the same agent working from the original papers and code. The reported benefit is quicker reuse of published methods and code. James Zou says paper-agents have helped him find potential collaborators; the article does not establish how broadly that result applies. Paper2Agent does not answer general questions about a paper, and the article reports that some papers could not be converted because of missing materials or software problems. Peter Clark raises a further concern: large numbers of low-quality paper-agents could pester scientists. The tests described do not establish whether paper-agents improve research outcomes or how well the approach works beyond the tested papers. How can researchers keep paper-agents useful and non-annoying as their numbers grow?
Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.
Original source
· Article
Source details
Original source date: Oct 21, 2026 (2026-10-21T13:00:00-04:00).
Date basis: Date stated by the publisher.
Added to Escape Meridian: 2026-11-14T09:16:25.556Z. First observed: 2026-10-21T17:04:22.021Z. Retrieved: 2026-10-22T06:20:09.329Z.
Referenced by
Recorded citations and public file attachments. A reference alone does not establish agreement or influence.
No public references recorded here yet.
Limits
The article reports tests on computational biology papers but does not establish how results generalize to other fields or research outcomes. Some papers could not be converted because of missing code or data, or software that would not run. The article does not quantify how often paper-agents identify useful collaborations or how they might affect scientists at scale.
Open question
How can researchers keep paper-agents useful and non-annoying as their numbers grow?
Source-stated
Paper2Agent packages a paper’s code, methods and data into tools for an AI assistant to use with new data.
When a paper comes with code, methods and data, the system packages these into a tool set that an AI assistant can use to work with new data
Source-stated
The researchers converted 74 of 100 computational biology papers into working tool sets.
tried to convert 100 computational biology papers into working tool sets and succeeded with 74.
Source-stated
For converted papers, Paper2Agent ran analyses and answered questions about results more accurately and more than twice as fast as the same agent working from the original papers and accompanying code.
It did so more accurately and also more than twice as fast as the same agent working from the original papers and accompanying code.
Source-stated
James Zou says paper-agents created from his team's work have helped him find other scientists to collaborate with.
Zou says his team has already created “over 100 different paper-agents” from their own work and that these have already helped him find other scientists to collaborate with.
Source-stated
Peter Clark says Paper2Agent focuses on reusing methods and code, not answering general questions about a paper.
Paper2Agent specializes in helping scientists reuse methods and code, and not in answering general questions, like a true virtual author should be able to do.
Source-stated
Clark worries that millions of low-quality paper-agents could pester scientists.
he worries about what could happen if millions of low-quality papers get turned into agents capable of pestering scientists.
Source and contributor context
Kathryn Hulick · Article author
These are credits stated by the source; they do not establish a profile match or authorship of every passage.
Source media
Media declared by the original publisher; availability checked separately from the editorial reading.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Science News
Science News reports scientific findings for a non-specialist reader.
Editorial provenance
Version 1. Publication ID: e1f9ead4-4ee0-4ada-8d22-8a006cf23b92.
Prepared by gpt-6-luna; grounding review by gpt-6-luna.
Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-14T09:16:25.556Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: material_development.
Observation: 7c8fd0b6bf569e7962b3a851a1c14935821b2f38b2342bb9342ed43dacde37e8. Source text hash: 75450e18bf3614252a8e08d582261770f3748f3d6e231f1b04103c023c91b2a6.
Revision and source-check history
Version 1 · 2026-11-14T09:16:25.556Z
Science News: Paper2Agent turns scientific papers into reusable AI tools
reports that Paper2Agent packages a scientific paper’s code, methods and data into tools an AI assistant can use with new data. The system also tests the tools against reference outputs and provides instructions for using them. In tests of 100 computational biology papers, researchers converted 74 into working tool sets. For converted papers, the system ran analyses and answered questions about results more accurately and more than twice as fast as the same agent working from the original papers and code. The reported benefit is quicker reuse of published methods and code. James Zou says paper-agents have helped him find potential collaborators; the article does not establish how broadly that result applies. Paper2Agent does not answer general questions about a paper, and the article reports that some papers could not be converted because of missing materials or software problems. Peter Clark raises a further concern: large numbers of low-quality paper-agents could pester scientists. The tests described do not establish whether paper-agents improve research outcomes or how well the approach works beyond the tested papers. How can researchers keep paper-agents useful and non-annoying as their numbers grow?
Original source
· Article
Source details
Original source date: Oct 21, 2026 (2026-10-21T13:00:00-04:00).
Date basis: Date stated by the publisher.
Added to Escape Meridian: 2026-11-14T09:16:25.556Z. First observed: 2026-10-21T17:04:22.021Z. Retrieved: 2026-10-22T06:20:09.329Z.
Limits
The article reports tests on computational biology papers but does not establish how results generalize to other fields or research outcomes. Some papers could not be converted because of missing code or data, or software that would not run. The article does not quantify how often paper-agents identify useful collaborations or how they might affect scientists at scale.
Open question
How can researchers keep paper-agents useful and non-annoying as their numbers grow?
Source-stated
Paper2Agent packages a paper’s code, methods and data into tools for an AI assistant to use with new data.
When a paper comes with code, methods and data, the system packages these into a tool set that an AI assistant can use to work with new data
Source-stated
The researchers converted 74 of 100 computational biology papers into working tool sets.
tried to convert 100 computational biology papers into working tool sets and succeeded with 74.
Source-stated
For converted papers, Paper2Agent ran analyses and answered questions about results more accurately and more than twice as fast as the same agent working from the original papers and accompanying code.
It did so more accurately and also more than twice as fast as the same agent working from the original papers and accompanying code.
Source-stated
James Zou says paper-agents created from his team's work have helped him find other scientists to collaborate with.
Zou says his team has already created “over 100 different paper-agents” from their own work and that these have already helped him find other scientists to collaborate with.
Source-stated
Peter Clark says Paper2Agent focuses on reusing methods and code, not answering general questions about a paper.
Paper2Agent specializes in helping scientists reuse methods and code, and not in answering general questions, like a true virtual author should be able to do.
Source-stated
Clark worries that millions of low-quality paper-agents could pester scientists.
he worries about what could happen if millions of low-quality papers get turned into agents capable of pestering scientists.
Source and contributor context
Kathryn Hulick · Article author
These are credits stated by the source; they do not establish a profile match or authorship of every passage.
Source media
Media declared by the original publisher; availability checked separately from the editorial reading.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Science News
Science News reports scientific findings for a non-specialist reader.
Editorial provenance
Version 1. Publication ID: e1f9ead4-4ee0-4ada-8d22-8a006cf23b92.
Prepared by gpt-6-luna; grounding review by gpt-6-luna.
Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-14T09:16:25.556Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: material_development.
Observation: 7c8fd0b6bf569e7962b3a851a1c14935821b2f38b2342bb9342ed43dacde37e8. Source text hash: 75450e18bf3614252a8e08d582261770f3748f3d6e231f1b04103c023c91b2a6.
uncertain · 2026-11-14T09:16:25.556Z · anthropic/claude-opus-5.5
Publication: e1f9ead4-4ee0-4ada-8d22-8a006cf23b92
attribution — uncertain: Most claims are attributed correctly to Science News, Zou or Clark. However, the summary sentence 'Paper2Agent does not answer general questions about a paper' presents Clark's assessment as unattributed fact. Smallest repair: 'Peter Clark notes that Paper2Agent specializes in reusing methods and code rather than answering general questions.' Zou and Clark are also not identified (Stanford computer scientist; Allen Institute researcher not involved in the work).
Clark notes that Paper2Agent specializes in helping scientists reuse methods and code, and not in answering general questions
summary — uncertain: The reading is accurate but omits key arguments and voices. Zou's thesis that papers should move from passive to active formats, his 'virtual author' framing, and Clark's view that this is 'an exciting step' are missing. Also missing is Clark's point that he would still write to authors about things left out of papers, which limits the virtual-author idea. The 'same agent' comparison and the actors' roles are not explained for a general reader. Smallest repair: add one sentence for Zou's argument and one for Clark's corresponding-author caveat.
now it’s time to move from passive to active formats for human knowledge
date — pass: originalPublished (2026-10-21T13:00-04:00) matches the Science News publisher page and is consistent with the feed entry. The Nature paper's September 16 date is not misrepresented as the article date.
support — uncertain: The figures (74 of 100; more accurate and more than twice as fast) and Zou's collaborator claim are source-stated. The flat 'does not answer general questions' is stronger than Clark's 'specializes in... not in answering general questions.' The limitations and question follow from Clark's concern and do not invent outcomes. Clark's concern is about low-quality papers turned into agents, which the draft renders closely enough.
he worries about what could happen if millions of low-quality papers get turned into agents capable of pestering scientists
media — pass: The selected media is the verified direct link to the original Science News article, the most direct way to access the source.
Challenge or correct this reading
Send a correction or an opposing source with this reading’s URL. Disagreement and uncertainty can remain unresolved.
Science News reports scientific findings for a non-specialist reader.
Explore Science NewsScience News reports scientific findings for a non-specialist reader.
Explore Science NewsScience News reports scientific findings for a non-specialist reader.
Explore Science NewsScience News reports scientific findings for a non-specialist reader.
Explore Science News