Can an AI search agent learn to look for evidence, not just get the answer right?
A peer-reviewed four-benchmark study found that rewarding useful search steps improved unfamiliar question answering and evidence retrieval. The tests used fixed academic datasets, GPT-4o annotations and extra inference compute—not live web research or outcomes for real users.
The short version
- What happened
- Researchers evaluated 3,398 questions across HotpotQA, 2WikiMultihopQA, Bamboogle and the English MedQA exam set, using different metrics for open-answer and multiple-choice tasks.
- Who it affects
- For clinicians, teachers, analysts and public-service teams, the practical lesson is to inspect the search trail as well as the answer. Process supervision may help, but consequential decisions still need source checks, domain review and a way to reject incomplete evidence.
- What remains unproven
- Fixed academic question-answering benchmarks rather than live web research
Keep in mind
Fixed academic question-answering benchmarks rather than live web research
No production users, decisions or harms measured
Most process labels generated by GPT-4o
Best-of-ten critic-guided inference adds compute and latency
Different benchmarks use different metrics and cannot be collapsed into one accuracy rate
No independent replication
English-language tasks dominate and MedQA is an exam set rather than clinical care
Earlier preprint existed before the 28 October peer-reviewed publication
Multi-source analysis · 3 linked sources. Primary analysed source date: 28 October 2026. Earliest linked supporting source: 16 May 2026.
Keep following the evidence
Join The Impact Brief founding list for selected AI reports, their effect on people and the evidence limits. Read a sample first.
Read the full report · 4 min The evidence, methods, implications and limitations
The agent was rewarded for how it searched
A newly peer-reviewed study suggests that information-seeking agents improve when training evaluates their intermediate searches, not only whether the final answer matches a key. That distinction matters for people using AI to assemble clinical, educational, policy or business evidence: a plausible answer can be right by luck or by relying on memorised model knowledge, while an auditable search process exposes what was actually retrieved.
The researchers built RAG-Gym, a common testbed for agent design, post-training and inference. Evaluation covered 3,398 questions: 1,000 each from HotpotQA and 2WikiMultihopQA, all 125 Bamboogle questions and 1,273 English MedQA questions. The first three are multi-step open-answer tasks scored with exact match and F1; MedQA is a multiple-choice medical exam set scored by accuracy. Separate samples of 1,000 HotpotQA and 1,000 MedQA training questions supplied process-reward data.[1][2]
Better benchmark scores came with useful evidence—and more machinery
The Re²Search agent drafts a line of reasoning, identifies the first unsupported claim and turns that gap into a targeted query. Its strongest version combines post-training with a critic that ranks candidate next actions. On HotpotQA, the full Re²Search++ system reached 60.19 F1, compared with 41.09 for a ReAct baseline. Across the study's comparisons with recent methods, the authors report average relative F1 gains of 3.2% to 11.6%, depending on the benchmark and configuration.
Some transfer results are encouraging but modest. A critic trained around Llama-3.1-8B improved GPT-4o-mini's MedQA accuracy from 76.8% to 78.1%. On 2Wiki, GPT-4o process annotations achieved 84.5% retrieval recall and supported 57.9 F1 downstream. The study therefore measured both answers and retrieved evidence—but did not ask independent experts whether the complete evidence trail would support a real decision.[1][2][3]
The distinction that matters
What changed in the training signal
The study separates a correct final answer from a search path that found relevant support.
Outcome only
Reward the final answer
A lucky or memorised answer may look successful
Process reward
Score each query and retrieved step
Useful evidence-seeking becomes trainable
Critic-guided
Sample ten next actions and rank them
Higher benchmark scores, with extra compute
The result is a research direction, not a reliability certificate
The critic normally chose among ten sampled actions at every step, increasing inference cost. More selection was not always better: MedQA accuracy fell from 72.8% with 15 candidates to 71.2% with 20, which the authors associate with reward hacking. An outcome-only method also beat one rollout-level process-reward variant on 2Wiki, 55.6 versus 49.6 F1. These counter-results make the central claim narrower than a universal win for process supervision.
GPT-4o generated most process labels. A medically trained author subset agreed with those labels 85.8% of the time on MedQA, but that is neither perfect agreement nor independent replication. The tasks were static benchmarks, not a changing web, long research session or live service. The work was funded by NIH and NSF awards; the authors declare no competing interests. Confidence would rise with independent replication, open-ended multilingual searches, cost and latency reporting, adversarial sources, calibrated citations and prospective tests measuring whether people make better decisions—not simply whether an answer key is matched.[1][2]
Watch the background, then check the evidence
Illustrative photo · not a video frame
Google DeepMind · 1 min 52 sec · Background explainer
Protein folding explained
YouTube loads only when you choose; playback starts when you press play.
A short introduction to why a protein’s shape matters. It provides background for understanding what AI structure prediction contributes to biological research.
Evidence check: This is the developer’s explanation. Predicting a structure does not by itself establish a treatment’s safety or benefit for patients.
Published 2020-11-30. Source checked 14 October 2026. Check captions on the publisher’s video page.
Explore current researchContinue the story
Read next
Continue exploring ai agents: compare the finding and its evidence limits.
Can clinical AI recognise when the patient record does not support an answer?
A new clinical-agent benchmark raises a practical question for health systems: can an assistant explain what the record cannot establish? Our analysis examines evidence, local testing and the burden on staff.
Read the next reportExplore the full source and evidence record
What this means for people
- For clinicians, teachers, analysts and public-service teams, the practical lesson is to inspect the search trail as well as the answer. Process supervision may help, but consequential decisions still need source checks, domain review and a way to reject incomplete evidence.
Global context
The authors are based at US institutions and the evaluated tasks are predominantly English-language benchmarks. Search quality will also depend on local-language coverage, access to primary sources, current data and the rules governing sensitive professional decisions in each jurisdiction.
What the evidence does not yet show
- Fixed academic question-answering benchmarks rather than live web research
- No production users, decisions or harms measured
- Most process labels generated by GPT-4o
- Best-of-ten critic-guided inference adds compute and latency
- Different benchmarks use different metrics and cannot be collapsed into one accuracy rate
- No independent replication
- English-language tasks dominate and MedQA is an exam set rather than clinical care
- Earlier preprint existed before the 28 October peer-reviewed publication
What to watch next
- Independent replication on live and changing sources
- Multilingual and region-specific retrieval
- Citation completeness and contradiction handling
- End-to-end cost, latency and energy use
- Performance under adversarial or low-quality sources
- Prospective human decision outcomes
- Whether smaller human-labelled datasets can reduce annotator-model dependence
Living evidence record
Impact record IAI-0CDCLYC
Evidence stage
Studied
Confidence
Corroborated
Reporting basis
Multi-source analysis
Independent or research support
Present
Record status
Monitoring
Last checked
28 October 2026
Source trail
3 direct sources across 2 source types.
People impact
Documented in this record.
Uncertainty
Limits and next checks are explicit.
Stages describe the evidence available—not whether a technology is good or bad. See the public method.
Evidence trail
Sources used for this report
Links checked 28 October 2026
Research question, topics and regions
Topics: AI agents · Information retrieval · Process supervision · Evidence quality · Question answering · Peer-reviewed research
Regions: United States · International
This report is labelled multi-source analysis. We summarise and analyse source material in our own words; company statements remain attributed claims until independently supported. Translated summaries preserve the meaning of the original source and link back to it. Read our editorial standards.
Reader commentary
What does this mean where you are?
Tell us what you have seen at work, in your community or in everyday life. A question is welcome too. No account is needed; we review notes before publishing them.
Explore commentary across the portal →Published reader notes
0No published reader notes yet. You can start the evidence-led discussion above.
Prefer a private correction or response? Contact the newsroom.