{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:VBEPQLJIPYCCRIU6BIXQDFBWNO","short_pith_number":"pith:VBEPQLJI","schema_version":"1.0","canonical_sha256":"a848f82d287e0428a29e0a2f0194366bb768f1f446ebc5d20974e9e4d024b17f","source":{"kind":"arxiv","id":"2605.20591","version":1},"attestation_state":"computed","paper":{"title":"Do No Harm? Hallucination and Actor-Level Abuse in Web-Deployed Medical Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CY"],"primary_cat":"cs.CL","authors_text":"Muhammad Ikram, Rahat Masood, Sunday Oyinlola Ogundoyin","submitted_at":"2026-05-20T00:57:59Z","abstract_excerpt":"Medical large language models (LLMs), including custom medical GPTs (MedGPTs) and open-source models, are increasingly deployed on web platforms to provide clinical guidance. However, they pose risks of hallucination, policy noncompliance, and unsafe design. We conduct a large-scale assessment of 6,233 MedGPTs, evaluating a stratified sample of 1,500, together with 10 open-source LLMs. We introduce two frameworks: MedGPT-HEval for hallucination detection and an LLM-based pipeline for assessing policy violations and developer intent. Our results show that 25-30% of MedGPTs exhibit low factual a"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2605.20591","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-20T00:57:59Z","cross_cats_sorted":["cs.CY"],"title_canon_sha256":"0cdc343a46ae4738162b373371fe52ea8b4725e26feb4c532b9b043d1836f9c3","abstract_canon_sha256":"be004539f3e13ab6e8d9740efe4d16edbb3d98cf3e813edb193a337ddd67bbac"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-21T01:04:43.787225Z","signature_b64":"hPtulTJxurAPU/lt5Ovqil39muWbKs4AMkv+r1o+nVvOD61G4MWkA3NGnchXBtsyf8783mn6ugTga7hQN3/RDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a848f82d287e0428a29e0a2f0194366bb768f1f446ebc5d20974e9e4d024b17f","last_reissued_at":"2026-05-21T01:04:43.786625Z","signature_status":"signed_v1","first_computed_at":"2026-05-21T01:04:43.786625Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Do No Harm? Hallucination and Actor-Level Abuse in Web-Deployed Medical Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CY"],"primary_cat":"cs.CL","authors_text":"Muhammad Ikram, Rahat Masood, Sunday Oyinlola Ogundoyin","submitted_at":"2026-05-20T00:57:59Z","abstract_excerpt":"Medical large language models (LLMs), including custom medical GPTs (MedGPTs) and open-source models, are increasingly deployed on web platforms to provide clinical guidance. However, they pose risks of hallucination, policy noncompliance, and unsafe design. We conduct a large-scale assessment of 6,233 MedGPTs, evaluating a stratified sample of 1,500, together with 10 open-source LLMs. We introduce two frameworks: MedGPT-HEval for hallucination detection and an LLM-based pipeline for assessing policy violations and developer intent. Our results show that 25-30% of MedGPTs exhibit low factual a"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.20591","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.20591/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2605.20591","created_at":"2026-05-21T01:04:43.786749+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.20591v1","created_at":"2026-05-21T01:04:43.786749+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.20591","created_at":"2026-05-21T01:04:43.786749+00:00"},{"alias_kind":"pith_short_12","alias_value":"VBEPQLJIPYCC","created_at":"2026-05-21T01:04:43.786749+00:00"},{"alias_kind":"pith_short_16","alias_value":"VBEPQLJIPYCCRIU6","created_at":"2026-05-21T01:04:43.786749+00:00"},{"alias_kind":"pith_short_8","alias_value":"VBEPQLJI","created_at":"2026-05-21T01:04:43.786749+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VBEPQLJIPYCCRIU6BIXQDFBWNO","json":"https://pith.science/pith/VBEPQLJIPYCCRIU6BIXQDFBWNO.json","graph_json":"https://pith.science/api/pith-number/VBEPQLJIPYCCRIU6BIXQDFBWNO/graph.json","events_json":"https://pith.science/api/pith-number/VBEPQLJIPYCCRIU6BIXQDFBWNO/events.json","paper":"https://pith.science/paper/VBEPQLJI"},"agent_actions":{"view_html":"https://pith.science/pith/VBEPQLJIPYCCRIU6BIXQDFBWNO","download_json":"https://pith.science/pith/VBEPQLJIPYCCRIU6BIXQDFBWNO.json","view_paper":"https://pith.science/paper/VBEPQLJI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.20591&json=true","fetch_graph":"https://pith.science/api/pith-number/VBEPQLJIPYCCRIU6BIXQDFBWNO/graph.json","fetch_events":"https://pith.science/api/pith-number/VBEPQLJIPYCCRIU6BIXQDFBWNO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VBEPQLJIPYCCRIU6BIXQDFBWNO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VBEPQLJIPYCCRIU6BIXQDFBWNO/action/storage_attestation","attest_author":"https://pith.science/pith/VBEPQLJIPYCCRIU6BIXQDFBWNO/action/author_attestation","sign_citation":"https://pith.science/pith/VBEPQLJIPYCCRIU6BIXQDFBWNO/action/citation_signature","submit_replication":"https://pith.science/pith/VBEPQLJIPYCCRIU6BIXQDFBWNO/action/replication_record"}},"created_at":"2026-05-21T01:04:43.786749+00:00","updated_at":"2026-05-21T01:04:43.786749+00:00"}