{ "id": "AIBD-0211", "title": "o3 Deep Research reanalysis at Boston Children's Hospital yields 18 confirmed local diagnoses, including 11 new diagnoses and seven rediscoveries", "summary": "AI-assisted reanalysis of 376 locally unresolved cases yielded 18 confirmed diagnoses. Eleven were new, while seven had already been established elsewhere.", "date_observed": "2026-06-18", "date_basis": "unassessed", "date_precision": "unassessed", "benefit_status": null, "source_collection": null, "source_entry": null, "account": "Boston Children’s Hospital and OpenAI researchers used o3 Deep Research to reassess 376 rare-disease cases that remained unresolved in their local research records. They reported 18 confirmed local diagnoses, or 4.8%: 11 newly identified diagnoses and seven rediscoveries of diagnoses already made elsewhere. Clinicians reviewed candidates, arranged laboratory confirmation, and returned results to families. The paper’s discussion explains that some findings depended on later gene discoveries, revised classifications, previously uncalled variants, or information that was unavailable to earlier analyses. The researchers did not compare the model with a simultaneous expert reanalysis using the same current evidence. They also did not measure improvements in treatment, cost, or clinician time. NBC News separately interviewed Kyra Benton about receiving a diagnosis and sought comment from outside clinicians. The study was retrospective and involved OpenAI employees and funding, so it does not establish an independent estimate of the model’s added benefit.", "category": "benefit created", "unexpected": false, "purpose": "OpenAI o3 Deep Research is a general-purpose reasoning and research model, not a diagnostic device. The Manton Center team built an explanation-first reanalysis workflow around it to propose candidate molecular diagnoses for unsolved cases. These diagnoses were the workflow's intended output. OpenAI states that the study does not describe or endorse a customer use of o3 for diagnosis. Brownstein said that she was shocked that a commercial system could find new diagnoses in genomes analyzed many times.", "system": "OpenAI o3 Deep Research, used as the reasoning layer of an LLM-assisted genome reanalysis workflow (Human Phenotype Ontology terms, clinician notes and a filtered variant table in; explanation-linked candidate diagnoses out)", "model_version": "o3 Deep Research", "developer": "OpenAI", "deployer": "Boston Children's Hospital, The Manton Center for Orphan Disease Research (with Harvard Medical School and OpenAI co-authors)", "location": "Boston, United States", "domain": "Clinical genomics / rare disease diagnosis", "segment": "Hospital rare-disease research center; retrospective reanalysis study", "benefit_type": "Improved Physical Health", "taxonomy_note": null, "registry_ids": { "oecd_aim": "2026-06-20-3da6" }, "ratings": [ { "group": "personal", "depth": 2, "reason": "Investigators returned 18 confirmed local diagnoses to families, including 11 new diagnoses and seven rediscoveries. NBC’s direct patient interview describes the value of receiving an explanation after years of uncertainty. Changes to treatment or health outcomes were not measured." }, { "group": "organizational", "depth": 1, "reason": "The team resolved 18 cases in its local research records, including seven diagnoses already established elsewhere. No contemporaneous human-reanalysis comparison or measured savings in time, cost, or clinician effort establish the additional organizational benefit of AI." } ], "ai_role": { "status": "assessed", "statement": "o3 Deep Research read each de-identified case packet and proposed the most plausible molecular explanation with its supporting evidence. It diagnosed no one. Humans did the rest. At least two reviewers adjudicated each candidate under ACMG/AMP criteria, a CLIA-certified laboratory confirmed each variant, and the clinical team returned results to families. Its distinct contribution was surfacing candidates in cases that commercial and institutional analyses and multidisciplinary teams had already reviewed. In one early-psychosis case it inferred a 22q11.2 deletion absent from the input data, which genome sequencing then confirmed. Seven of the 18 were rediscoveries already established externally. The study was retrospective, and reviewers were not blinded to model confidence. The authors call for prospective evaluation with comparator baselines, so the gain over renewed human reanalysis is unmeasured." }, "evidence": [ { "kind": "peer-reviewed study", "citation": "Jaech A, Cheatham M, Shringarpure SS, Genetti CA, Pradhan P, Bagul A, Panch T, Rader B, Sewalk K, Distler R, Anderson KN, Stewart M, Lejfer S, Glahn DC, Goldstein RD, Wojcik M, Beggs AH, Brownstein JS, Brownstein CA (2026). LLM-Assisted Reanalysis of Unsolved Rare Disease Genomes Increases Diagnostic Yield. NEJM AI 3(7); doi:10.1056/AIcs2501343; PMID 42529462. Four of the nineteen authors, including the first, are OpenAI employees; the senior author is at Boston Children's Hospital and Harvard Medical School.", "title": "LLM-Assisted Reanalysis of Unsolved Rare Disease Genomes Increases Diagnostic Yield", "url": "https://doi.org/10.1056/AIcs2501343", "date": "2026-06-25", "written_by": "makers", "quote": null }, { "kind": "news report", "citation": "Jared Perlo. “AI helped diagnose 18 children whose rare diseases had stumped doctors.” NBC News, June 18, 2026. Original interviews with patient Kyra Benton, study investigators, and outside clinicians.", "title": "AI helped diagnose 18 children whose rare diseases had stumped doctors", "url": "https://www.nbcnews.com/tech/innovation/ai-boston-childrens-hospital-diagnose-rare-diseases-kids-openai-rcna350387", "date": "2026-06-18", "written_by": "independent", "quote": null } ], "replication": { "produced_by": [ "Already-sequenced unsolved cases, mostly child-parent trios, assembled into a de-identified packet. Each packet held Human Phenotype Ontology terms, any clinician notes, age and sex. It also held a filtered variant table with rarity, predicted effect, ClinVar class and family signal quality", "A general-purpose reasoning model asked for the single most plausible molecular explanation and its evidence, not just a ranked gene. The prompt was refined first on cases of known diagnosis", "Time elapsed since the original analysis, so that newly published gene-disease links and reclassified variants existed to be found", "Expert adjudication under ACMG/AMP criteria by at least two reviewers, CLIA-laboratory confirmation, and return of results by the clinical team" ], "must_hold": [ "Human geneticists review every candidate and a model output is never treated as a diagnosis", "Confirmatory testing in a certified laboratory before any result reaches a family", "De-identified data handled inside approved environments", "Yield expectations calibrated to the cohort: single-digit gains in heavily reviewed cases, and seven of the 18 here were rediscoveries" ], "echoes": [] }, "paired_incident": null, "headline": "o3-assisted reanalysis yields 11 new diagnoses and seven rediscoveries", "metric": "gave 18 families confirmed rare-disease diagnoses", "comparator": { "against": "prior local research records; no contemporaneous expert-reanalysis control", "before": "376 cases unresolved in the local research record, including some diagnosed elsewhere", "after": "18 confirmed local diagnoses: 11 new and seven rediscoveries; 4.8% overall", "basis": "measured", "quote": "We identified seven rediscoveries", "url": "https://doi.org/10.1056/AIcs2501343" }, "later": { "status": "unknown", "date": null, "note": "The paper calls for prospective multicenter evaluation, but no later development after the June 2026 publication was found.", "quote": null, "url": null }, "value": null, "developer_country": "United States", "deployer_country": "United States", "unrated_because": null, "brief_notes": null, "brief": null, "revisions": [ { "date": "2026-10-20", "change": "Record drafted from the OECD Monitor entry and the cited sources.", "source": "https://oecd.ai/en/incidents/2026-06-20-3da6", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Benefit type set to Improved Physical Health under Trustible's full taxonomy of 23 types; it was Improved Knowledge Discovery among the six first used.", "source": "https://trustible.ai/post/a-framework-for-measuring-the-benefits-of-ai/", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Added what the benefit is measured against from the cited sources.", "source": "https://europepmc.org/abstract/MED/42529462", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Country of the deployer added from the organization's own site or a registry.", "source": "https://en.wikipedia.org/wiki/Boston_Children%27s_Hospital", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Country of the developer added from the organization's own site or a registry.", "source": "https://en.wikipedia.org/wiki/OpenAI", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Prose pass: a rating reason, account, purpose, the replication conditions, the system's part rewritten in shorter sentences with abbreviations expanded; facts, figures and quotations unchanged.", "source": null, "editor": "AIBD" }, { "date": "2026-10-24", "change": "Read the full paper, corrected 18 wholly new diagnoses to 11 new plus seven rediscoveries, and retained NBC for original patient and outside-expert interviews.", "source": "https://doi.org/10.1056/AIcs2501343", "editor": "AIBD" }, { "date": "2026-10-25", "change": "Clarified the displayed field answers and placed measurements and limitations in their respective fields.", "source": null, "editor": "AIBD" } ], "confidence": "reported", "shade": "medium", "claim_status": "Independently reported", "claim_basis": "Sources: one independent news report and the makers' peer-reviewed study. The cited evidence includes no independent study or report and no finding by a regulator or court.", "profile": { "case_id": "AIBD-0211", "catalog_version": "1.0.0", "source_fingerprint": "62db4c890827eaf5d1a8644a78979a919343bda8e166a386481a64749fd0aa91", "collection_refs": [ { "collection_id": "aibd", "version": "snapshot-8bf7cfec9cb185aa" }, { "collection_id": "aibd", "version": "2026-10-25" } ], "source_ids": [ "AIBD-0211-source-0001", "AIBD-0211-source-0002" ], "families": { "strategic assessments": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "record_ids": [] }, "evaluations": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "record_ids": [] }, "monetary observations": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "record_ids": [] }, "institutional decisions": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "record_ids": [] } }, "fields": { "actors[]": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "developer": "OpenAI", "deployer": "Boston Children's Hospital, The Manton Center for Orphan Disease Research (with Harvard Medical School and OpenAI co-authors)" } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "developer", "deployer" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "actors[].ownership_control": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "ai_contribution": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "ai_role": { "status": "assessed", "statement": "o3 Deep Research read each de-identified case packet and proposed the most plausible molecular explanation with its supporting evidence. It diagnosed no one. Humans did the rest. At least two reviewers adjudicated each candidate under ACMG/AMP criteria, a CLIA-certified laboratory confirmed each variant, and the clinical team returned results to families. Its distinct contribution was surfacing candidates in cases that commercial and institutional analyses and multidisciplinary teams had already reviewed. In one early-psychosis case it inferred a 22q11.2 deletion absent from the input data, which genome sequencing then confirmed. Seven of the 18 were rediscoveries already established externally. The study was retrospective, and reviewers were not blinded to model confidence. The authors call for prospective evaluation with comparator baselines, so the gain over renewed human reanalysis is unmeasured." } } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "ai_role" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "ai_relationship": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "category": "benefit created", "ai_role": { "status": "assessed", "statement": "o3 Deep Research read each de-identified case packet and proposed the most plausible molecular explanation with its supporting evidence. It diagnosed no one. Humans did the rest. At least two reviewers adjudicated each candidate under ACMG/AMP criteria, a CLIA-certified laboratory confirmed each variant, and the clinical team returned results to families. Its distinct contribution was surfacing candidates in cases that commercial and institutional analyses and multidisciplinary teams had already reviewed. In one early-psychosis case it inferred a 22q11.2 deletion absent from the input data, which genome sequencing then confirmed. Seven of the 18 were rediscoveries already established externally. The study was retrospective, and reviewers were not blinded to model confidence. The authors call for prospective evaluation with comparator baselines, so the gain over renewed human reanalysis is unmeasured." } } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "ai_role", "category" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "assessments[].alternatives": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "assessments[].assessor_date": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "assessments[].depth": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "ratings": [ { "group": "personal", "depth": 2, "reason": "Investigators returned 18 confirmed local diagnoses to families, including 11 new diagnoses and seven rediscoveries. NBC’s direct patient interview describes the value of receiving an explanation after years of uncertainty. Changes to treatment or health outcomes were not measured." }, { "group": "organizational", "depth": 1, "reason": "The team resolved 18 cases in its local research records, including seven diagnoses already established elsewhere. No contemporaneous human-reanalysis comparison or measured savings in time, cost, or clinician effort establish the additional organizational benefit of AI." } ] } }, "provenance": { "kind": "curatorial", "basis": "Updated affected case fields after the dated primary-source review. Unrelated requirement assessments retain their existing status.", "record_fields": [ "ratings" ], "source_refs": [ "AIBD-0211-source-0001", "AIBD-0211-source-0002" ], "reviewed_on": "2026-10-24" }, "review_state": "record_mapped", "coverage": "partial" }, "assessments[].display_confidence": { "state": "value_present", "value": { "kind": "structured", "data": { "confidence": "medium", "evidence_word": "reported", "claim_status": "Independently reported", "basis": "Sources: one independent news report and the makers' peer-reviewed study. The cited evidence includes no independent study or report and no finding by a regulator or court.", "interpretation": "Existing display rule preserved; the richer outcome-specific confidence question still needs review." } }, "provenance": { "kind": "curatorial", "basis": "Derived from the existing case display properties; not a new evidentiary judgment.", "record_fields": [], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "assessments[].rationale": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "ratings": [ { "group": "personal", "depth": 2, "reason": "Investigators returned 18 confirmed local diagnoses to families, including 11 new diagnoses and seven rediscoveries. NBC’s direct patient interview describes the value of receiving an explanation after years of uncertainty. Changes to treatment or health outcomes were not measured." }, { "group": "organizational", "depth": 1, "reason": "The team resolved 18 cases in its local research records, including seven diagnoses already established elsewhere. No contemporaneous human-reanalysis comparison or measured savings in time, cost, or clinician effort establish the additional organizational benefit of AI." } ] } }, "provenance": { "kind": "curatorial", "basis": "Updated affected case fields after the dated primary-source review. Unrelated requirement assessments retain their existing status.", "record_fields": [ "ratings" ], "source_refs": [ "AIBD-0211-source-0001", "AIBD-0211-source-0002" ], "reviewed_on": "2026-10-24" }, "review_state": "record_mapped", "coverage": "partial", "display": { "answer": "Families received 18 confirmed local diagnoses, including 11 new findings. A patient interview describes the value of receiving an explanation. The study measured no changes in health or treatment, and no comparison isolates AI’s added value over an expert reanalysis." } }, "assessments[].rubric_version": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "assessments[].status_basis": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "assessments[].target": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "ratings": [ { "group": "personal", "depth": 2, "reason": "Investigators returned 18 confirmed local diagnoses to families, including 11 new diagnoses and seven rediscoveries. NBC’s direct patient interview describes the value of receiving an explanation after years of uncertainty. Changes to treatment or health outcomes were not measured." }, { "group": "organizational", "depth": 1, "reason": "The team resolved 18 cases in its local research records, including seven diagnoses already established elsewhere. No contemporaneous human-reanalysis comparison or measured savings in time, cost, or clinician effort establish the additional organizational benefit of AI." } ] } }, "provenance": { "kind": "curatorial", "basis": "Updated affected case fields after the dated primary-source review. Unrelated requirement assessments retain their existing status.", "record_fields": [ "ratings" ], "source_refs": [ "AIBD-0211-source-0001", "AIBD-0211-source-0002" ], "reviewed_on": "2026-10-24" }, "review_state": "record_mapped", "coverage": "partial" }, "availability": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "brief": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "case_kind": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "case_links[]": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "claims[].evidence_links": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "claims[].evidence_strength": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "claims[].missingness": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "claims[].proposition": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "cross_border_flows[]": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "decisions[].action_outcome": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "institutional decisions" }, "decisions[].grounds": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "institutional decisions" }, "decisions[].institution_task": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "institutional decisions" }, "decisions[].success_criterion": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "institutional decisions" }, "deployment_stage": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "economics[].amount": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "monetary observations" }, "economics[].costs[]": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "monetary observations" }, "economics[].counterfactual": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "monetary observations" }, "economics[].currency_price_basis": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "monetary observations" }, "economics[].gross_net": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "monetary observations" }, "economics[].measure_type": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "monetary observations" }, "economics[].period_unit": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "monetary observations" }, "economics[].realization": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "monetary observations" }, "economics[].recipient_payer": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "monetary observations" }, "economics[].return_payback": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "monetary observations" }, "economics[].valuation_method": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "monetary observations" }, "evaluations[].ai_uplift": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "evaluations" }, "evaluations[].comparator": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "evaluations" }, "evaluations[].design": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "evaluations" }, "evaluations[].effect": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "evaluations" }, "evaluations[].exposure": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "evaluations" }, "evaluations[].id": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "evaluations" }, "evaluations[].measurement_method": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "evaluations", "display": { "answer": "Experts reviewed the proposed diagnoses under ACMG/AMP criteria and arranged laboratory confirmation. Researchers counted confirmed diagnoses against 376 locally unresolved cases." } }, "evaluations[].metric": { "state": "not_assessed", "provenance": { "kind": "legacy_record", "basis": "Mapped from the existing case description without a new external source review.", "record_fields": [ "metric" ], "source_refs": [] }, "review_state": "not_assessed", "coverage": "partial", "family": "evaluations", "display": { "answer": "Confirmed diagnoses, distinguishing new diagnoses from those already established elsewhere." } }, "evaluations[].reproduction_materials": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "evaluations" }, "evaluations[].setting_sample": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "evaluations", "display": { "answer": "A retrospective reanalysis of 376 rare-disease records at Boston Children’s Hospital, mostly involving a child and both parents." } }, "evaluations[].uncertainty_limitations": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "evaluations", "display": { "answer": "There was no simultaneous expert reanalysis using the same updated evidence. Some findings relied on later gene discoveries or revised variant classifications. The study did not measure treatment improvements, savings, or clinician time, and involved OpenAI employees and funding." } }, "external_references[]": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "registry_ids": { "oecd_aim": "2026-06-20-3da6" } } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "registry_ids" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "follow_up[]": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "later": { "status": "unknown", "note": "The paper calls for prospective multicenter evaluation, but no later development after the June 2026 publication was found." } } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "later" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "human_oversight": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "id": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "id": "AIBD-0211" } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "id" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "infrastructure_dependencies[]": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "outcomes[].beneficiaries": { "state": "value_present", "value": { "kind": "text", "text": "Families whose unresolved local records were reexamined at Boston Children’s Hospital." }, "provenance": { "kind": "legacy_record", "basis": "Moved the specific information already present in the entry into its applicable question. No new source review is implied.", "record_fields": [ "account" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial", "display": { "answer": "Families whose unresolved local records were reexamined at Boston Children’s Hospital." } }, "outcomes[].beneficiary_count": { "state": "value_present", "value": { "kind": "text", "text": "Eighteen families received confirmed local diagnoses, including 11 newly diagnosed families." }, "provenance": { "kind": "legacy_record", "basis": "Moved the specific information already present in the entry into its applicable question. No new source review is implied.", "record_fields": [ "account", "metric" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial", "display": { "answer": "Eighteen families received confirmed local diagnoses, including 11 newly diagnosed families." } }, "outcomes[].benefit_type": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "benefit_type": "Improved Physical Health" } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "benefit_type" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "outcomes[].claim": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "account": "Boston Children’s Hospital and OpenAI researchers used o3 Deep Research to reassess 376 rare-disease cases that remained unresolved in their local research records. They reported 18 confirmed local diagnoses, or 4.8%: 11 newly identified diagnoses and seven rediscoveries of diagnoses already made elsewhere. Clinicians reviewed candidates, arranged laboratory confirmation, and returned results to families. The paper’s discussion explains that some findings depended on later gene discoveries, revised classifications, previously uncalled variants, or information that was unavailable to earlier analyses. The researchers did not compare the model with a simultaneous expert reanalysis using the same current evidence. They also did not measure improvements in treatment, cost, or clinician time. NBC News separately interviewed Kyra Benton about receiving a diagnosis and sought comment from outside clinicians. The study was retrospective and involved OpenAI employees and funding, so it does not establish an independent estimate of the model’s added benefit.", "benefit_type": "Improved Physical Health" } }, "provenance": { "kind": "curatorial", "basis": "Updated affected case fields after the dated primary-source review. Unrelated requirement assessments retain their existing status.", "record_fields": [ "account", "benefit_type" ], "source_refs": [ "AIBD-0211-source-0001", "AIBD-0211-source-0002" ], "reviewed_on": "2026-10-24" }, "review_state": "record_mapped", "coverage": "partial", "display": { "answer": "Researchers confirmed 18 diagnoses in local records: 11 new diagnoses and seven already established elsewhere." } }, "outcomes[].distribution": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "outcomes[].duration": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "outcomes[].id": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "outcomes[].stakeholder_group": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "ratings": [ { "group": "personal", "depth": 2, "reason": "Investigators returned 18 confirmed local diagnoses to families, including 11 new diagnoses and seven rediscoveries. NBC’s direct patient interview describes the value of receiving an explanation after years of uncertainty. Changes to treatment or health outcomes were not measured." }, { "group": "organizational", "depth": 1, "reason": "The team resolved 18 cases in its local research records, including seven diagnoses already established elsewhere. No contemporaneous human-reanalysis comparison or measured savings in time, cost, or clinician effort establish the additional organizational benefit of AI." } ] } }, "provenance": { "kind": "curatorial", "basis": "Updated affected case fields after the dated primary-source review. Unrelated requirement assessments retain their existing status.", "record_fields": [ "ratings" ], "source_refs": [ "AIBD-0211-source-0001", "AIBD-0211-source-0002" ], "reviewed_on": "2026-10-24" }, "review_state": "record_mapped", "coverage": "partial" }, "outcomes[].status": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "outcomes[].tradeoffs": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "outcomes[].unexpected": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "unexpected": false, "purpose": "OpenAI o3 Deep Research is a general-purpose reasoning and research model, not a diagnostic device. The Manton Center team built an explanation-first reanalysis workflow around it to propose candidate molecular diagnoses for unsolved cases. These diagnoses were the workflow's intended output. OpenAI states that the study does not describe or endorse a customer use of o3 for diagnosis. Brownstein said that she was shocked that a commercial system could find new diagnoses in genomes analyzed many times." } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "unexpected", "purpose" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "places[]": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "location": "Boston, United States", "developer_country": "United States", "deployer_country": "United States" } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "location", "developer_country", "deployer_country" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "purpose": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "purpose": "OpenAI o3 Deep Research is a general-purpose reasoning and research model, not a diagnostic device. The Manton Center team built an explanation-first reanalysis workflow around it to propose candidate molecular diagnoses for unsolved cases. These diagnoses were the workflow's intended output. OpenAI states that the study does not describe or endorse a customer use of o3 for diagnosis. Brownstein said that she was shocked that a commercial system could find new diagnoses in genomes analyzed many times." } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "purpose" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "record_dates": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "date_observed": "2026-06-18", "revisions": [ { "date": "2026-10-20", "change": "Record drafted from the OECD Monitor entry and the cited sources.", "source": "https://oecd.ai/en/incidents/2026-06-20-3da6", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Benefit type set to Improved Physical Health under Trustible's full taxonomy of 23 types; it was Improved Knowledge Discovery among the six first used.", "source": "https://trustible.ai/post/a-framework-for-measuring-the-benefits-of-ai/", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Added what the benefit is measured against from the cited sources.", "source": "https://europepmc.org/abstract/MED/42529462", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Country of the deployer added from the organization's own site or a registry.", "source": "https://en.wikipedia.org/wiki/Boston_Children%27s_Hospital", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Country of the developer added from the organization's own site or a registry.", "source": "https://en.wikipedia.org/wiki/OpenAI", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Prose pass: a rating reason, account, purpose, the replication conditions, the system's part rewritten in shorter sentences with abbreviations expanded; facts, figures and quotations unchanged.", "editor": "AIBD" }, { "date": "2026-10-24", "change": "Read the full paper, corrected 18 wholly new diagnoses to 11 new plus seven rediscoveries, and retained NBC for original patient and outside-expert interviews.", "source": "https://doi.org/10.1056/AIcs2501343", "editor": "AIBD" }, { "date": "2026-10-25", "change": "Clarified the displayed field answers and placed measurements and limitations in their respective fields.", "editor": "AIBD" } ] } }, "provenance": { "kind": "curatorial", "basis": "Updated affected case fields after the dated primary-source review. Unrelated requirement assessments retain their existing status.", "record_fields": [ "date_observed", "revisions" ], "source_refs": [ "AIBD-0211-source-0001", "AIBD-0211-source-0002" ], "reviewed_on": "2026-10-24" }, "review_state": "record_mapped", "coverage": "partial" }, "replication.attempts[]": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "replication.conditions": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "replication": { "produced_by": [ "Already-sequenced unsolved cases, mostly child-parent trios, assembled into a de-identified packet. Each packet held Human Phenotype Ontology terms, any clinician notes, age and sex. It also held a filtered variant table with rarity, predicted effect, ClinVar class and family signal quality", "A general-purpose reasoning model asked for the single most plausible molecular explanation and its evidence, not just a ranked gene. The prompt was refined first on cases of known diagnosis", "Time elapsed since the original analysis, so that newly published gene-disease links and reclassified variants existed to be found", "Expert adjudication under ACMG/AMP criteria by at least two reviewers, CLIA-laboratory confirmation, and return of results by the clinical team" ], "must_hold": [ "Human geneticists review every candidate and a model output is never treated as a diagnosis", "Confirmatory testing in a certified laboratory before any result reaches a family", "De-identified data handled inside approved environments", "Yield expectations calibrated to the cohort: single-digit gains in heavily reviewed cases, and seven of the 18 here were rediscoveries" ], "echoes": [] } } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "replication" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial", "display": { "answer": "Geneticists must review candidates and arrange confirmatory tests. The data need approved, protected handling. Expected yield depends on the prior work and the newer evidence available for each cohort." } }, "revision_history[]": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "revisions": [ { "date": "2026-10-20", "change": "Record drafted from the OECD Monitor entry and the cited sources.", "source": "https://oecd.ai/en/incidents/2026-06-20-3da6", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Benefit type set to Improved Physical Health under Trustible's full taxonomy of 23 types; it was Improved Knowledge Discovery among the six first used.", "source": "https://trustible.ai/post/a-framework-for-measuring-the-benefits-of-ai/", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Added what the benefit is measured against from the cited sources.", "source": "https://europepmc.org/abstract/MED/42529462", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Country of the deployer added from the organization's own site or a registry.", "source": "https://en.wikipedia.org/wiki/Boston_Children%27s_Hospital", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Country of the developer added from the organization's own site or a registry.", "source": "https://en.wikipedia.org/wiki/OpenAI", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Prose pass: a rating reason, account, purpose, the replication conditions, the system's part rewritten in shorter sentences with abbreviations expanded; facts, figures and quotations unchanged.", "editor": "AIBD" }, { "date": "2026-10-24", "change": "Read the full paper, corrected 18 wholly new diagnoses to 11 new plus seven rediscoveries, and retained NBC for original patient and outside-expert interviews.", "source": "https://doi.org/10.1056/AIcs2501343", "editor": "AIBD" }, { "date": "2026-10-25", "change": "Clarified the displayed field answers and placed measurements and limitations in their respective fields.", "editor": "AIBD" } ] } }, "provenance": { "kind": "curatorial", "basis": "Updated affected case fields after the dated primary-source review. Unrelated requirement assessments retain their existing status.", "record_fields": [ "revisions" ], "source_refs": [ "AIBD-0211-source-0001", "AIBD-0211-source-0002" ], "reviewed_on": "2026-10-24" }, "review_state": "record_mapped", "coverage": "partial" }, "sector": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "domain": "Clinical genomics / rare disease diagnosis", "segment": "Hospital rare-disease research center; retrospective reanalysis study" } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "domain", "segment" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "sources[].access_version": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "sources[].citation_locator": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "evidence": [ { "kind": "peer-reviewed study", "citation": "Jaech A, Cheatham M, Shringarpure SS, Genetti CA, Pradhan P, Bagul A, Panch T, Rader B, Sewalk K, Distler R, Anderson KN, Stewart M, Lejfer S, Glahn DC, Goldstein RD, Wojcik M, Beggs AH, Brownstein JS, Brownstein CA (2026). LLM-Assisted Reanalysis of Unsolved Rare Disease Genomes Increases Diagnostic Yield. NEJM AI 3(7); doi:10.1056/AIcs2501343; PMID 42529462. Four of the nineteen authors, including the first, are OpenAI employees; the senior author is at Boston Children's Hospital and Harvard Medical School.", "title": "LLM-Assisted Reanalysis of Unsolved Rare Disease Genomes Increases Diagnostic Yield", "url": "https://doi.org/10.1056/AIcs2501343", "date": "2026-06-25", "written_by": "makers" }, { "kind": "news report", "citation": "Jared Perlo. “AI helped diagnose 18 children whose rare diseases had stumped doctors.” NBC News, June 18, 2026. Original interviews with patient Kyra Benton, study investigators, and outside clinicians.", "title": "AI helped diagnose 18 children whose rare diseases had stumped doctors", "url": "https://www.nbcnews.com/tech/innovation/ai-boston-childrens-hospital-diagnose-rare-diseases-kids-openai-rcna350387", "date": "2026-06-18", "written_by": "independent" } ] } }, "provenance": { "kind": "source_review", "basis": "Read the full paper, corrected 18 wholly new diagnoses to 11 new plus seven rediscoveries, and retained NBC for original patient and outside-expert interviews.", "record_fields": [ "evidence" ], "source_refs": [ "AIBD-0211-source-0001", "AIBD-0211-source-0002" ], "reviewed_on": "2026-10-24" }, "review_state": "reviewed", "coverage": "partial" }, "sources[].id": { "state": "value_present", "value": { "kind": "identifiers", "identifiers": [ "AIBD-0211-source-0001", "AIBD-0211-source-0002" ], "role": "Active evidence sources in reviewed order" }, "provenance": { "kind": "curatorial", "basis": "Assigned to the retained evidence in the dated source review.", "record_fields": [ "evidence" ], "source_refs": [ "AIBD-0211-source-0001", "AIBD-0211-source-0002" ], "reviewed_on": "2026-10-24" }, "review_state": "reviewed", "coverage": "complete" }, "sources[].language": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "sources[].relationship": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "evidence": [ { "kind": "peer-reviewed study", "citation": "Jaech A, Cheatham M, Shringarpure SS, Genetti CA, Pradhan P, Bagul A, Panch T, Rader B, Sewalk K, Distler R, Anderson KN, Stewart M, Lejfer S, Glahn DC, Goldstein RD, Wojcik M, Beggs AH, Brownstein JS, Brownstein CA (2026). LLM-Assisted Reanalysis of Unsolved Rare Disease Genomes Increases Diagnostic Yield. NEJM AI 3(7); doi:10.1056/AIcs2501343; PMID 42529462. Four of the nineteen authors, including the first, are OpenAI employees; the senior author is at Boston Children's Hospital and Harvard Medical School.", "title": "LLM-Assisted Reanalysis of Unsolved Rare Disease Genomes Increases Diagnostic Yield", "url": "https://doi.org/10.1056/AIcs2501343", "date": "2026-06-25", "written_by": "makers" }, { "kind": "news report", "citation": "Jared Perlo. “AI helped diagnose 18 children whose rare diseases had stumped doctors.” NBC News, June 18, 2026. Original interviews with patient Kyra Benton, study investigators, and outside clinicians.", "title": "AI helped diagnose 18 children whose rare diseases had stumped doctors", "url": "https://www.nbcnews.com/tech/innovation/ai-boston-childrens-hospital-diagnose-rare-diseases-kids-openai-rcna350387", "date": "2026-06-18", "written_by": "independent" } ] } }, "provenance": { "kind": "source_review", "basis": "Read the full paper, corrected 18 wholly new diagnoses to 11 new plus seven rediscoveries, and retained NBC for original patient and outside-expert interviews.", "record_fields": [ "evidence" ], "source_refs": [ "AIBD-0211-source-0001", "AIBD-0211-source-0002" ], "reviewed_on": "2026-10-24" }, "review_state": "reviewed", "coverage": "partial" }, "sources[].upstream_sources": { "state": "value_present", "value": { "kind": "structured", "data": { "origins": [ { "source_refs": [ "AIBD-0211-source-0001" ], "url": "https://doi.org/10.1056/AIcs2501343", "title": "LLM-Assisted Reanalysis of Unsolved Rare Disease Genomes Increases Diagnostic Yield", "published_on": "2026-06-25", "relationship": "makers", "citation": "Jaech A, Cheatham M, Shringarpure SS, Genetti CA, Pradhan P, Bagul A, Panch T, Rader B, Sewalk K, Distler R, Anderson KN, Stewart M, Lejfer S, Glahn DC, Goldstein RD, Wojcik M, Beggs AH, Brownstein JS, Brownstein CA (2026). LLM-Assisted Reanalysis of Unsolved Rare Disease Genomes Increases Diagnostic Yield. NEJM AI 3(7); doi:10.1056/AIcs2501343; PMID 42529462. Four of the nineteen authors, including the first, are OpenAI employees; the senior author is at Boston Children's Hospital and Harvard Medical School." }, { "source_refs": [ "AIBD-0211-source-0002" ], "url": "https://www.nbcnews.com/tech/innovation/ai-boston-childrens-hospital-diagnose-rare-diseases-kids-openai-rcna350387", "title": "AI helped diagnose 18 children whose rare diseases had stumped doctors", "published_on": "2026-06-18", "relationship": "independent", "citation": "Jared Perlo. “AI helped diagnose 18 children whose rare diseases had stumped doctors.” NBC News, June 18, 2026. Original interviews with patient Kyra Benton, study investigators, and outside clinicians." } ], "excluded_republications": [ { "url": "https://openai.com/index/diagnose-rare-childhood-diseases/", "title": "Using AI to help physicians diagnose rare genetic diseases affecting children", "relationship": "Retained as source history; replaced by the original evidence or omitted because no distinct evidence is used for the revised claims", "previous_source_ref": "AIBD-0211-source-0003" }, { "url": "https://exame.com/inteligencia-artificial/depois-de-anos-sem-respostas-ia-ajuda-hospital-a-diagnosticar-doencas-raras/", "title": "Depois de anos sem respostas, IA ajuda hospital a diagnosticar doenças raras", "published_on": "2026-06-24", "relationship": "Retained as source history; replaced by the original evidence or omitted because no distinct evidence is used for the revised claims", "previous_source_ref": "AIBD-0211-source-0004" }, { "url": "https://olhardigital.com.br/2026/06/21/medicina-e-saude/ia-openai-acelera-diagnostico-de-doencas-raras-em-hospital-dos-eua/", "title": "IA OpenAI acelera diagnóstico de doenças raras em hospital dos EUA", "published_on": "2026-06-21", "relationship": "Retained as source history; replaced by the original evidence or omitted because no distinct evidence is used for the revised claims", "previous_source_ref": "AIBD-0211-source-0005" }, { "url": "https://europepmc.org/abstract/MED/42529462", "title": "LLM-Assisted Reanalysis of Unsolved Rare Disease Genomes Increases Diagnostic Yield (Europe PMC abstract)", "published_on": "2026-06-18", "relationship": "Retained as source history; replaced by the original evidence or omitted because no distinct evidence is used for the revised claims", "previous_source_ref": "AIBD-0211-source-0006" } ], "independent_evaluations_of_reported_results": 0, "lead_source_ref": "AIBD-0211-source-0001", "retained_source_order": [ "AIBD-0211-source-0001", "AIBD-0211-source-0002" ], "ordering_basis": "The source that directly supports the central case leads. Additional sources contribute the specific evidence or limitations described in claim_lineage.", "claim_lineage": [ { "claim": "Diagnostic yield and clinical confirmation", "source_chain": [ { "url": "https://doi.org/10.1056/AIcs2501343", "locator": "Results, Table 1, Discussion, and funding/conflict declarations", "role": "Original peer-reviewed study", "access_status": "Full article reviewed through NCBI XML https://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=pmc&id=13417654", "published_on": "2026-06-25" }, { "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC13417654/", "locator": "Supplementary Appendix 1 (NIHMS2191980-supplement-appendix_1.pdf) and Appendix 2 (NIHMS2191980-supplement-appendix_2.xlsx)", "role": "Methods/prompts and patient/benchmark supporting files", "access_status": "File names identified in full JATS; attempted downloads failed" } ], "limitations": [ "11 newly identified diagnoses and seven rediscoveries; no contemporaneous expert-reanalysis comparator", "Several cases reflect later discoveries or revised interpretation rather than an established AI-specific advantage", "Supplementary files and disclosure PDF were not retrieved" ], "source_ref": "AIBD-0211-source-0001" }, { "claim": "Patient experience and independent clinical comment", "source_chain": [ { "url": "https://www.nbcnews.com/tech/innovation/ai-boston-childrens-hospital-diagnose-rare-diseases-kids-openai-rcna350387", "locator": "Kyra Benton interview; comments from Adam Rodman and Chunhua Weng", "role": "Original reporting and interviews", "access_status": "Full relevant text reviewed", "published_on": "2026-06-18" } ], "limitations": [ "Original patient testimony and independent commentary are not replication of the study" ], "source_ref": "AIBD-0211-source-0002" }, { "claim": "Publication chronology and discrepant biographical detail", "source_chain": [ { "url": "https://www.nbcnews.com/tech/innovation/ai-boston-childrens-hospital-diagnose-rare-diseases-kids-openai-rcna350387", "locator": "First report of the study and patient interview", "role": "Original dated news report", "access_status": "Full relevant text reviewed", "published_on": "2026-06-18" }, { "url": "https://openai.com/index/diagnose-rare-childhood-diseases/", "locator": "Dated research-partner announcement", "role": "Original developer announcement", "access_status": "Full relevant text reviewed", "published_on": "2026-06-18" }, { "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC13417654/", "locator": "JATS article-meta publication date", "role": "Journal publication metadata", "access_status": "Full metadata reviewed; epub date is June 25, 2026" } ], "limitations": [ "June 18 announcements precede the June 25 electronic publication date in NCBI metadata; the record retains June 18 as first public reporting", "NBC quotes the patient’s twentieth birthday, whereas OpenAI’s account says twenty-eighth; the revised headline omits the unresolved age" ], "source_ref": "AIBD-0211-source-0002" } ], "limitations": [ "Supplements and disclosure PDF remain inaccessible" ], "chronology_source": { "url": "https://www.nbcnews.com/tech/innovation/ai-boston-childrens-hospital-diagnose-rare-diseases-kids-openai-rcna350387", "locator": "First verified public report", "role": "Earliest dated account of this clinical result", "access_status": "Full relevant text reviewed", "published_on": "2026-06-18", "date": "2026-06-18", "title": "AI helped diagnose 18 children whose rare diseases had stumped doctors" } } }, "provenance": { "kind": "source_review", "basis": "Read the full paper, corrected 18 wholly new diagnoses to 11 new plus seven rediscoveries, and retained NBC for original patient and outside-expert interviews.", "record_fields": [ "evidence" ], "source_refs": [ "AIBD-0211-source-0001", "AIBD-0211-source-0002" ], "reviewed_on": "2026-10-24" }, "review_state": "reviewed", "coverage": "partial", "display": { "answer": "The study’s supplements and disclosure document were inaccessible.", "editor_note": "" } }, "strategic_assessments[]": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "strategic assessments" }, "summary": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "summary": "AI-assisted reanalysis of 376 locally unresolved cases yielded 18 confirmed diagnoses. Eleven were new, while seven had already been established elsewhere.", "account": "Boston Children’s Hospital and OpenAI researchers used o3 Deep Research to reassess 376 rare-disease cases that remained unresolved in their local research records. They reported 18 confirmed local diagnoses, or 4.8%: 11 newly identified diagnoses and seven rediscoveries of diagnoses already made elsewhere. Clinicians reviewed candidates, arranged laboratory confirmation, and returned results to families. The paper’s discussion explains that some findings depended on later gene discoveries, revised classifications, previously uncalled variants, or information that was unavailable to earlier analyses. The researchers did not compare the model with a simultaneous expert reanalysis using the same current evidence. They also did not measure improvements in treatment, cost, or clinician time. NBC News separately interviewed Kyra Benton about receiving a diagnosis and sought comment from outside clinicians. The study was retrospective and involved OpenAI employees and funding, so it does not establish an independent estimate of the model’s added benefit." } }, "provenance": { "kind": "curatorial", "basis": "Updated affected case fields after the dated primary-source review. Unrelated requirement assessments retain their existing status.", "record_fields": [ "summary", "account" ], "source_refs": [ "AIBD-0211-source-0001", "AIBD-0211-source-0002" ], "reviewed_on": "2026-10-24" }, "review_state": "record_mapped", "coverage": "partial" }, "systems[].autonomy_permissions": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "systems[].component_role": { "state": "value_present", "value": { "kind": "text", "text": "The model proposed molecular explanations from clinical information and genetic variants." }, "provenance": { "kind": "legacy_record", "basis": "Restated the mechanical task already described in the existing system and AI-role fields. No new source review is claimed.", "record_fields": [ "system", "ai_role" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "systems[].deployment_configuration": { "state": "value_present", "value": { "kind": "text", "text": "Researchers used o3 Deep Research to analyze de-identified clinical information and genetic variants." }, "provenance": { "kind": "legacy_record", "basis": "Separated the model, software, hardware, or operation already described in the existing record. No new source review is claimed.", "record_fields": [ "model_version", "system", "ai_role" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "systems[].inputs_outputs": { "state": "value_present", "value": { "kind": "text", "text": "The model received clinical notes, phenotype terms, and filtered genetic variants. It proposed a molecular diagnosis with supporting evidence." }, "provenance": { "kind": "legacy_record", "basis": "Separated the model, software, hardware, or operation already described in the existing record. No new source review is claimed.", "record_fields": [ "system", "ai_role", "replication", "account" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "systems[].model_name": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "systems[].model_version": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "model_version": "o3 Deep Research" } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "model_version" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial", "display": { "answer": "Exact release not established.", "editor_note": "" } }, "systems[].product_name": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "system": "OpenAI o3 Deep Research, used as the reasoning layer of an LLM-assisted genome reanalysis workflow (Human Phenotype Ontology terms, clinician notes and a filtered variant table in; explanation-linked candidate diagnoses out)" } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "system" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "systems[].provider": { "state": "value_present", "value": { "kind": "structured", "data": { "application_developer": "OpenAI" } }, "provenance": { "kind": "legacy_record", "basis": "Mapped from the existing case description without a new external source review.", "record_fields": [ "developer" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "taxonomy_crosswalks[]": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "title": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "title": "o3 Deep Research reanalysis at Boston Children's Hospital yields 18 confirmed local diagnoses, including 11 new diagnoses and seven rediscoveries", "headline": "o3-assisted reanalysis yields 11 new diagnoses and seven rediscoveries" } }, "provenance": { "kind": "curatorial", "basis": "Updated affected case fields after the dated primary-source review. Unrelated requirement assessments retain their existing status.", "record_fields": [ "title", "headline" ], "source_refs": [ "AIBD-0211-source-0001", "AIBD-0211-source-0002" ], "reviewed_on": "2026-10-24" }, "review_state": "record_mapped", "coverage": "partial" }, "version": { "state": "value_present", "value": { "kind": "identifiers", "identifiers": [ "62db4c890827eaf5d1a8644a78979a919343bda8e166a386481a64749fd0aa91" ], "role": "SHA-256 of normalized original case data" }, "provenance": { "kind": "curatorial", "basis": "Deterministic content version after the October 24, 2026 primary-source review.", "record_fields": [], "source_refs": [], "reviewed_on": "2026-10-24" }, "review_state": "record_mapped", "coverage": "complete" }, "workflow": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "account": "Boston Children’s Hospital and OpenAI researchers used o3 Deep Research to reassess 376 rare-disease cases that remained unresolved in their local research records. They reported 18 confirmed local diagnoses, or 4.8%: 11 newly identified diagnoses and seven rediscoveries of diagnoses already made elsewhere. Clinicians reviewed candidates, arranged laboratory confirmation, and returned results to families. The paper’s discussion explains that some findings depended on later gene discoveries, revised classifications, previously uncalled variants, or information that was unavailable to earlier analyses. The researchers did not compare the model with a simultaneous expert reanalysis using the same current evidence. They also did not measure improvements in treatment, cost, or clinician time. NBC News separately interviewed Kyra Benton about receiving a diagnosis and sought comment from outside clinicians. The study was retrospective and involved OpenAI employees and funding, so it does not establish an independent estimate of the model’s added benefit.", "replication": { "produced_by": [ "Already-sequenced unsolved cases, mostly child-parent trios, assembled into a de-identified packet. Each packet held Human Phenotype Ontology terms, any clinician notes, age and sex. It also held a filtered variant table with rarity, predicted effect, ClinVar class and family signal quality", "A general-purpose reasoning model asked for the single most plausible molecular explanation and its evidence, not just a ranked gene. The prompt was refined first on cases of known diagnosis", "Time elapsed since the original analysis, so that newly published gene-disease links and reclassified variants existed to be found", "Expert adjudication under ACMG/AMP criteria by at least two reviewers, CLIA-laboratory confirmation, and return of results by the clinical team" ], "must_hold": [ "Human geneticists review every candidate and a model output is never treated as a diagnosis", "Confirmatory testing in a certified laboratory before any result reaches a family", "De-identified data handled inside approved environments", "Yield expectations calibrated to the cohort: single-digit gains in heavily reviewed cases, and seven of the 18 here were rediscoveries" ], "echoes": [] } } }, "provenance": { "kind": "curatorial", "basis": "Updated affected case fields after the dated primary-source review. Unrelated requirement assessments retain their existing status.", "record_fields": [ "account", "replication" ], "source_refs": [ "AIBD-0211-source-0001", "AIBD-0211-source-0002" ], "reviewed_on": "2026-10-24" }, "review_state": "record_mapped", "coverage": "partial", "display": { "answer": "Researchers gave the model de-identified symptoms, clinical notes, and genetic variants that earlier analyses had not resolved. They refined the prompt on known diagnoses, then asked for explanations of unresolved cases. At least two experts reviewed each candidate, and a certified laboratory confirmed diagnoses before clinicians returned results to families." } } }, "summary": { "total": 91, "addressed": 91, "known": 36, "reviewed": 4, "record_mapped": 32, "not_assessed": 55, "complete": false, "by_state": { "value_present": 36, "not_reported": 0, "unresolved": 0, "not_assessed": 55, "not_applicable": 0, "withheld": 0, "source_unavailable": 0 } } } }