{"id":"https://openalex.org/W7166543788","doi":"https://doi.org/10.1145/3805689.3812332","title":"Expert Evaluation and the Limits of Human Feedback in Mental Health AI Safety Testing","display_name":"Expert Evaluation and the Limits of Human Feedback in Mental Health AI Safety Testing","publication_year":2026,"publication_date":"2026-06-25","ids":{"openalex":"https://openalex.org/W7166543788","doi":"https://doi.org/10.1145/3805689.3812332"},"language":null,"primary_location":{"id":"doi:10.1145/3805689.3812332","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3805689.3812332","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2026 ACM Conference on Fairness, Accountability, and Transparency","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1145/3805689.3812332","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5081841950","display_name":"Kiana Jafari Meimandi","orcid":"https://orcid.org/0000-0002-2001-3488"},"institutions":[{"id":"https://openalex.org/I97018004","display_name":"Stanford University","ror":"https://ror.org/00f54p054","country_code":"US","type":"education","lineage":["https://openalex.org/I97018004"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Kiana Jafari Meimandi","raw_affiliation_strings":["Stanford University, Stanford, CA, USA"],"raw_orcid":"https://orcid.org/0000-0002-2001-3488","affiliations":[{"raw_affiliation_string":"Stanford University, Stanford, CA, USA","institution_ids":["https://openalex.org/I97018004"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5123890267","display_name":"Paul Ulrich Nikolaus Rust","orcid":null},"institutions":[{"id":"https://openalex.org/I97018004","display_name":"Stanford University","ror":"https://ror.org/00f54p054","country_code":"US","type":"education","lineage":["https://openalex.org/I97018004"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Paul Ulrich Nikolaus Rust","raw_affiliation_strings":["Stanford University, Stanford, CA, USA"],"raw_orcid":"https://orcid.org/0009-0000-8375-5617","affiliations":[{"raw_affiliation_string":"Stanford University, Stanford, CA, USA","institution_ids":["https://openalex.org/I97018004"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5077282337","display_name":"Duncan Eddy","orcid":null},"institutions":[{"id":"https://openalex.org/I97018004","display_name":"Stanford University","ror":"https://ror.org/00f54p054","country_code":"US","type":"education","lineage":["https://openalex.org/I97018004"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Duncan Eddy","raw_affiliation_strings":["Stanford University, Stanford, CA, USA"],"raw_orcid":"https://orcid.org/0009-0000-2832-9711","affiliations":[{"raw_affiliation_string":"Stanford University, Stanford, CA, USA","institution_ids":["https://openalex.org/I97018004"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5124048064","display_name":"Robbie Fraser","orcid":null},"institutions":[{"id":"https://openalex.org/I45129253","display_name":"University College London","ror":"https://ror.org/02jx3x895","country_code":"GB","type":"education","lineage":["https://openalex.org/I124357947","https://openalex.org/I45129253"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Robbie Fraser","raw_affiliation_strings":["University College London, London, United Kingdom"],"raw_orcid":"https://orcid.org/0009-0007-2277-0785","affiliations":[{"raw_affiliation_string":"University College London, London, United Kingdom","institution_ids":["https://openalex.org/I45129253"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5027136972","display_name":"Nina Vasan","orcid":"https://orcid.org/0000-0002-8917-9320"},"institutions":[{"id":"https://openalex.org/I97018004","display_name":"Stanford University","ror":"https://ror.org/00f54p054","country_code":"US","type":"education","lineage":["https://openalex.org/I97018004"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Nina Vasan","raw_affiliation_strings":["Stanford University, Stanford, CA, USA"],"raw_orcid":"https://orcid.org/0000-0002-8917-9320","affiliations":[{"raw_affiliation_string":"Stanford University, Stanford, CA, USA","institution_ids":["https://openalex.org/I97018004"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5017512752","display_name":"Darja Djordjevic","orcid":"https://orcid.org/0000-0002-7012-6755"},"institutions":[{"id":"https://openalex.org/I2801851002","display_name":"Harvard University Press","ror":"https://ror.org/006v7bf86","country_code":"US","type":"other","lineage":["https://openalex.org/I136199984","https://openalex.org/I2801851002"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Darja Djordjevic","raw_affiliation_strings":["Darja@post.harvard.edu, Cambridge, MA, USA"],"raw_orcid":"https://orcid.org/0000-0002-7012-6755","affiliations":[{"raw_affiliation_string":"Darja@post.harvard.edu, Cambridge, MA, USA","institution_ids":["https://openalex.org/I2801851002"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5051046030","display_name":"Akanksha Dadlani","orcid":"https://orcid.org/0000-0002-3595-3173"},"institutions":[{"id":"https://openalex.org/I97018004","display_name":"Stanford University","ror":"https://ror.org/00f54p054","country_code":"US","type":"education","lineage":["https://openalex.org/I97018004"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Akanksha Dadlani","raw_affiliation_strings":["Stanford University, Stanford, CA, USA"],"raw_orcid":"https://orcid.org/0000-0002-3595-3173","affiliations":[{"raw_affiliation_string":"Stanford University, Stanford, CA, USA","institution_ids":["https://openalex.org/I97018004"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5090231346","display_name":"Max Lamparth","orcid":"https://orcid.org/0000-0001-6405-513X"},"institutions":[{"id":"https://openalex.org/I97018004","display_name":"Stanford University","ror":"https://ror.org/00f54p054","country_code":"US","type":"education","lineage":["https://openalex.org/I97018004"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Max Lamparth","raw_affiliation_strings":["Stanford University, Stanford, CA, USA"],"raw_orcid":"https://orcid.org/0000-0001-6405-513X","affiliations":[{"raw_affiliation_string":"Stanford University, Stanford, CA, USA","institution_ids":["https://openalex.org/I97018004"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5047872416","display_name":"Eugenia Kim","orcid":"https://orcid.org/0009-0003-9619-0768"},"institutions":[{"id":"https://openalex.org/I1290206253","display_name":"Microsoft (United States)","ror":"https://ror.org/00d0nc645","country_code":"US","type":"company","lineage":["https://openalex.org/I1290206253"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Eugenia Kim","raw_affiliation_strings":["Microsoft, Redmond, WA, USA"],"raw_orcid":"https://orcid.org/0009-0003-9619-0768","affiliations":[{"raw_affiliation_string":"Microsoft, Redmond, WA, USA","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5068326377","display_name":"Mykel J. Kochenderfer","orcid":"https://orcid.org/0000-0002-7238-9663"},"institutions":[{"id":"https://openalex.org/I97018004","display_name":"Stanford University","ror":"https://ror.org/00f54p054","country_code":"US","type":"education","lineage":["https://openalex.org/I97018004"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Mykel Kochenderfer","raw_affiliation_strings":["Stanford University, Stanford, CA, USA"],"raw_orcid":"https://orcid.org/0000-0002-7238-9663","affiliations":[{"raw_affiliation_string":"Stanford University, Stanford, CA, USA","institution_ids":["https://openalex.org/I97018004"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":4,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"3894","last_page":"3920"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.15219999849796295,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.15219999849796295,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11636","display_name":"Artificial Intelligence in Healthcare and Education","score":0.14380000531673431,"subfield":{"id":"https://openalex.org/subfields/2718","display_name":"Health Informatics"},"field":{"id":"https://openalex.org/fields/27","display_name":"Medicine"},"domain":{"id":"https://openalex.org/domains/4","display_name":"Health Sciences"}},{"id":"https://openalex.org/T10883","display_name":"Ethics and Social Impacts of AI","score":0.04859999939799309,"subfield":{"id":"https://openalex.org/subfields/3311","display_name":"Safety Research"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/mental-health","display_name":"Mental health","score":0.5480999946594238},{"id":"https://openalex.org/keywords/expert-system","display_name":"Expert system","score":0.33640000224113464},{"id":"https://openalex.org/keywords/control","display_name":"Control (management)","score":0.2858999967575073},{"id":"https://openalex.org/keywords/medline","display_name":"MEDLINE","score":0.2802000045776367},{"id":"https://openalex.org/keywords/risk-assessment","display_name":"Risk assessment","score":0.2793999910354614},{"id":"https://openalex.org/keywords/test","display_name":"Test (biology)","score":0.2745000123977661}],"concepts":[{"id":"https://openalex.org/C134362201","wikidata":"https://www.wikidata.org/wiki/Q317309","display_name":"Mental health","level":2,"score":0.5480999946594238},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.5019000172615051},{"id":"https://openalex.org/C75630572","wikidata":"https://www.wikidata.org/wiki/Q538904","display_name":"Applied psychology","level":1,"score":0.4625999927520752},{"id":"https://openalex.org/C71924100","wikidata":"https://www.wikidata.org/wiki/Q11190","display_name":"Medicine","level":0,"score":0.3806000053882599},{"id":"https://openalex.org/C509550671","wikidata":"https://www.wikidata.org/wiki/Q126945","display_name":"Medical education","level":1,"score":0.36570000648498535},{"id":"https://openalex.org/C58328972","wikidata":"https://www.wikidata.org/wiki/Q184609","display_name":"Expert system","level":2,"score":0.33640000224113464},{"id":"https://openalex.org/C118552586","wikidata":"https://www.wikidata.org/wiki/Q7867","display_name":"Psychiatry","level":1,"score":0.28850001096725464},{"id":"https://openalex.org/C112930515","wikidata":"https://www.wikidata.org/wiki/Q4389547","display_name":"Risk analysis (engineering)","level":1,"score":0.28610000014305115},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.28600001335144043},{"id":"https://openalex.org/C2775924081","wikidata":"https://www.wikidata.org/wiki/Q55608371","display_name":"Control (management)","level":2,"score":0.2858999967575073},{"id":"https://openalex.org/C2779473830","wikidata":"https://www.wikidata.org/wiki/Q1540899","display_name":"MEDLINE","level":2,"score":0.2802000045776367},{"id":"https://openalex.org/C12174686","wikidata":"https://www.wikidata.org/wiki/Q1058438","display_name":"Risk assessment","level":2,"score":0.2793999910354614},{"id":"https://openalex.org/C159110408","wikidata":"https://www.wikidata.org/wiki/Q121176","display_name":"Nursing","level":1,"score":0.27459999918937683},{"id":"https://openalex.org/C2777267654","wikidata":"https://www.wikidata.org/wiki/Q3519023","display_name":"Test (biology)","level":2,"score":0.2745000123977661},{"id":"https://openalex.org/C70410870","wikidata":"https://www.wikidata.org/wiki/Q199906","display_name":"Clinical psychology","level":1,"score":0.26759999990463257},{"id":"https://openalex.org/C2780665704","wikidata":"https://www.wikidata.org/wiki/Q959298","display_name":"Intervention (counseling)","level":2,"score":0.2671000063419342},{"id":"https://openalex.org/C77805123","wikidata":"https://www.wikidata.org/wiki/Q161272","display_name":"Social psychology","level":1,"score":0.2669000029563904},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.2535000145435333},{"id":"https://openalex.org/C2776035688","wikidata":"https://www.wikidata.org/wiki/Q1606558","display_name":"Affect (linguistics)","level":2,"score":0.2524999976158142}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3805689.3812332","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3805689.3812332","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2026 ACM Conference on Fairness, Accountability, and Transparency","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3805689.3812332","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3805689.3812332","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2026 ACM Conference on Fairness, Accountability, and Transparency","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":53,"referenced_works":["https://openalex.org/W2402700","https://openalex.org/W1962580118","https://openalex.org/W1968564639","https://openalex.org/W1979290264","https://openalex.org/W2012378416","https://openalex.org/W2062236692","https://openalex.org/W2063085086","https://openalex.org/W2067724039","https://openalex.org/W2118347738","https://openalex.org/W2119461458","https://openalex.org/W2123822033","https://openalex.org/W2152462550","https://openalex.org/W2167659819","https://openalex.org/W2181251292","https://openalex.org/W2271992345","https://openalex.org/W2410895653","https://openalex.org/W2476805942","https://openalex.org/W2623779865","https://openalex.org/W2895995041","https://openalex.org/W2935892262","https://openalex.org/W2985347336","https://openalex.org/W3104982372","https://openalex.org/W3126938010","https://openalex.org/W3197587259","https://openalex.org/W4225006216","https://openalex.org/W4226404919","https://openalex.org/W4234180827","https://openalex.org/W4247665917","https://openalex.org/W4287674181","https://openalex.org/W4311991106","https://openalex.org/W4317757464","https://openalex.org/W4378468563","https://openalex.org/W4378771755","https://openalex.org/W4380353763","https://openalex.org/W4385436553","https://openalex.org/W4387223260","https://openalex.org/W4389713766","https://openalex.org/W4391177205","https://openalex.org/W4394567566","https://openalex.org/W4394875731","https://openalex.org/W4396985403","https://openalex.org/W4401042481","https://openalex.org/W4404783621","https://openalex.org/W4406152279","https://openalex.org/W4408084770","https://openalex.org/W4410032125","https://openalex.org/W4411542009","https://openalex.org/W4412944817","https://openalex.org/W4414801979","https://openalex.org/W7102388716","https://openalex.org/W7133189643","https://openalex.org/W7133196894","https://openalex.org/W7133211372"],"related_works":[],"abstract_inverted_index":{"Learning":[0],"from":[1,195,205],"human":[2],"feedback":[3],"(LHF)":[4],"assumes":[5],"that":[6,111,131,146,153,192,201],"expert":[7,33,162,206],"judgments,":[8],"appropriately":[9],"aggregated,":[10],"yield":[11],"valid":[12],"ground":[13],"truth":[14],"for":[15,65,183],"training":[16,49],"and":[17,50,77,87,122,188,203],"evaluating":[18],"AI":[19,166],"systems.":[20],"We":[21,180],"tested":[22],"this":[23],"assumption":[24],"in":[25,164],"mental":[26],"health,":[27],"where":[28,171],"high":[29],"safety":[30,186],"stakes":[31],"make":[32],"consensus":[34],"essential.":[35],"Three":[36],"certified":[37],"psychiatrists":[38],"independently":[39],"evaluated":[40],"LLM-generated":[41],"responses":[42,79],"using":[43],"a":[44,99,168],"calibrated":[45],"rubric.":[46],"Despite":[47],"similar":[48],"shared":[51],"instructions,":[52],"inter-rater":[53],"reliability":[54,97],"was":[55,69,88],"consistently":[56],"poor":[57],"(ICC":[58],"0.087-0.295),":[59],"falling":[60],"below":[61],"thresholds":[62],"considered":[63],"acceptable":[64],"consequential":[66],"assessment.":[67],"Disagreement":[68],"highest":[70],"on":[71,134],"the":[72],"most":[73],"safety-critical":[74,165],"items.":[75],"Suicide":[76],"self-harm":[78],"produced":[80],"greater":[81],"divergence":[82],"than":[83,91,106,126,139],"any":[84],"other":[85],"category,":[86],"systematic":[89],"rather":[90,125,138],"random.":[92],"One":[93],"factor":[94,141],"yielded":[95],"negative":[96],"(Krippendorff's":[98],"=":[100],"-0.203),":[101],"indicating":[102],"structured":[103],"disagreement":[104,112,163],"worse":[105],"chance.":[107],"Qualitative":[108],"interviews":[109],"revealed":[110],"reflects":[113],"coherent":[114],"but":[115],"incompatible":[116],"individual":[117],"clinical":[118],"frameworks,":[119],"safety-first,":[120],"engagement-centered,":[121],"culturally-informed":[123],"orientations,":[124],"measurement":[127],"error.":[128],"By":[129],"demonstrating":[130],"experts":[132],"rely":[133],"holistic":[135],"risk":[136],"heuristics":[137],"granular":[140],"discrimination,":[142],"these":[143],"findings":[144],"suggest":[145],"aggregated":[147],"labels":[148],"function":[149],"as":[150,167],"arithmetic":[151],"compromises":[152],"effectively":[154],"erase":[155],"grounded":[156],"professional":[157,172],"philosophies.":[158],"Our":[159],"results":[160],"characterize":[161],"sociotechnical":[169],"phenomenon":[170],"experience":[173],"introduces":[174],"sophisticated":[175],"layers":[176],"of":[177],"principled":[178],"divergence.":[179],"discuss":[181],"implications":[182],"reward":[184],"modeling,":[185],"classification,":[187],"evaluation":[189],"benchmarks,":[190],"recommending":[191],"practitioners":[193],"shift":[194],"consensus-based":[196],"aggregation":[197],"to":[198],"alignment":[199],"methods":[200],"preserve":[202],"learn":[204],"disagreement.":[207]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2026-06-30T00:00:00"}