Would the same argument sway an AI differently depending on who benefits?
FairStress tests 300 two-candidate decisions across 29 domains and finds modest baseline accuracy gaps but larger asymmetries when one sentence pressures a model to change a correct answer. The scenarios are synthetic, not field outcomes.
The short version
- What happened
- FairStress starts with 300 synthetic decisions across 29 domains, renders more than 13 million combinations and provides a 50,012-item core set selected with item-response theory.
- Who it affects
- Organisations using AI to rank or recommend people should test not only whether baseline outputs look balanced, but whether identical evidence and pressure change decisions differently across groups. Human review and appeal remain essential.
- What remains unproven
- The cases are synthetic, short and limited to two candidates.
FairStress tests both the answer and how firmly it is held
The team created 300 two-person decisions across 29 areas including hiring, housing, healthcare and law. Each appears with an objectively stronger candidate or a genuine tie, 33 demographic pairings, swapped answer order and four identity levels from none to an explicit label. Combining these with 78 one-sentence prompts produced 13,032,104 items; a 50,012-item core subset was selected for lower-cost testing.[1][2]
A benchmark asymmetry is not a finding about real people
The scenarios are short, synthetic and US-centred. Pressure is one sentence, the training-data audit covers only OLMo-3, and activation analysis uses two models and indirect probes. The result supports testing whether AI decisions remain stable when social pressure changes, rather than assuming a fair-looking answer reflects a fair process. Field audits, broader cultural settings and independent replication would be needed before drawing conclusions about deployed systems.[1][2]
Who it affects
Organisations using AI to rank or recommend people should test not only whether baseline outputs look balanced, but whether identical evidence and pressure change decisions differently across groups. Human review and appeal remain essential.[1][2]
Keep in mind
The cases are synthetic, short and limited to two candidates.
The demographic convention and equality definition are US-centred and do not cover every conception of fairness.
One-sentence pressure does not represent every system prompt or multi-turn interaction.
Activation and reasoning probes are indirect; the authors describe their separability as small.
The institute report has not yet been independently replicated in deployed systems.
Multi-source analysis · 2 linked sources. Original source date: 20 October 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 · 3 min The evidence, methods, implications and limitations
FairStress tests both the answer and how firmly it is held
The team created 300 two-person decisions across 29 areas including hiring, housing, healthcare and law. Each appears with an objectively stronger candidate or a genuine tie, 33 demographic pairings, swapped answer order and four identity levels from none to an explicit label. Combining these with 78 one-sentence prompts produced 13,032,104 items; a 50,012-item core subset was selected for lower-cost testing.[1][2]
Pressure exposed a larger asymmetry than baseline accuracy
Across 15 models on the full corpus, explicit identity produced a mean 2.9-percentage-point accuracy gap. When an added sentence argued for the weaker candidate, the models abandoned correct answers favouring the majority-coded person 9.3 points more often. In tied cases, explicit labels led models to choose the minority-coded candidate 61.6% of the time on average rather than the 50% equality reference. Human checks agreed with 94% of 100 ground-truth pairs and called 90% of 100 ambiguous pairs ties.[1]
The distinction that matters
Four layers of one decision
FairStress asks whether equal treatment survives beyond the first answer.
Decision
Compare choices with matched facts
Measure the visible answer
Pressure
Add a matched persuasive sentence
Measure asymmetric switching
Reasoning
Compare explanations and activations
Look for process differences
A benchmark asymmetry is not a finding about real people
The scenarios are short, synthetic and US-centred. Pressure is one sentence, the training-data audit covers only OLMo-3, and activation analysis uses two models and indirect probes. The result supports testing whether AI decisions remain stable when social pressure changes, rather than assuming a fair-looking answer reflects a fair process. Field audits, broader cultural settings and independent replication would be needed before drawing conclusions about deployed systems.[1][2]
Watch the background, then check the evidence
Illustrative photo · not a video frame
The Royal Society · 44 min 28 sec · Background explainer
Is this the AI future we want? | The Royal Society
YouTube loads only when you choose; playback starts when you press play.
A Royal Society discussion with Madhumita Murgia and biomedical engineer Alison Noble on AI’s potential in healthcare, science and culture, recorded at the 30 June 2026 event.
Evidence check: A discussion of opportunities and risks, including future possibilities. It is not evidence that every application described is available or clinically proven.
Published 2026-07-01. Source checked 14 October 2026. Check captions on the publisher’s video page.
Check the healthcare evidencePhoto: Daniel Sone / National Cancer Institute / Unsplash · Licence
Evidence check
The cases are synthetic, short and limited to two candidates.
The demographic convention and equality definition are US-centred and do not cover every conception of fairness.
One-sentence pressure does not represent every system prompt or multi-turn interaction.
Activation and reasoning probes are indirect; the authors describe their separability as small.
The institute report has not yet been independently replicated in deployed systems.
Reporting basis: Multi-source analysis. Original source date: 20 October 2026.
Continue the story
Read next
Continue exploring benchmarking: compare the finding and its evidence limits.
Can a robot infer what a person intends?
A peer-reviewed video-language model matched or exceeded reported human scores on five-choice intention questions. When answer options disappeared, text-overlap scores fell below 20—leaving open-vocabulary claims, cultural bias and surveillance risk unresolved.
Read the next reportExplore the full source and evidence record
What this means for people
- Organisations using AI to rank or recommend people should test not only whether baseline outputs look balanced, but whether identical evidence and pressure change decisions differently across groups. Human review and appeal remain essential.
Global context
Decision systems cross legal and cultural boundaries where protected groups, fairness duties and acceptable trade-offs differ. Local governance and appeal mechanisms cannot be replaced by one benchmark score.
What the evidence does not yet show
- The cases are synthetic, short and limited to two candidates.
- The demographic convention and equality definition are US-centred and do not cover every conception of fairness.
- One-sentence pressure does not represent every system prompt or multi-turn interaction.
- Activation and reasoning probes are indirect; the authors describe their separability as small.
- The institute report has not yet been independently replicated in deployed systems.
What to watch next
- Independent replication
- Field audits in deployed systems
- Cross-cultural scenarios
- Interventions that reduce asymmetry without destroying accuracy
Living evidence record
Impact record IAI-0V66QJZ
Evidence stage
Studied
Confidence
Supported
Reporting basis
Multi-source analysis
Independent or research support
Not yet
Record status
Monitoring
Last checked
21 October 2026
Source trail
2 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 21 October 2026
Research question, topics and regions
Whether language models make and defend equivalent two-person decisions symmetrically when demographic identity and persuasive pressure change
Topics: Fairness · Decision systems · Demographic bias · Model robustness · Benchmarking
Regions: United States · Global
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.