{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:XS2AMAZCKELXARJNATVFQWN6WR","short_pith_number":"pith:XS2AMAZC","schema_version":"1.0","canonical_sha256":"bcb4060322511770452d04ea5859beb478e94824867fcbfba46ef2063b0fd5a9","source":{"kind":"arxiv","id":"2501.13302","version":1},"attestation_state":"computed","paper":{"title":"Watching the AI Watchdogs: A Fairness and Robustness Analysis of AI Safety Moderation Classifiers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Akshit Achara, Anshuman Chhabra","submitted_at":"2025-01-23T01:04:00Z","abstract_excerpt":"AI Safety Moderation (ASM) classifiers are designed to moderate content on social media platforms and to serve as guardrails that prevent Large Language Models (LLMs) from being fine-tuned on unsafe inputs. Owing to their potential for disparate impact, it is crucial to ensure that these classifiers: (1) do not unfairly classify content belonging to users from minority groups as unsafe compared to those from majority groups and (2) that their behavior remains robust and consistent across similar inputs. In this work, we thus examine the fairness and robustness of four widely-used, closed-sourc"},"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":"2501.13302","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-01-23T01:04:00Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"cb823df72f2432da0232d1abe4933204e3f0a491a1e98a4a0393428f6b090f5d","abstract_canon_sha256":"f78e49abef854157b29eb90680c1cb37fde08186e5559afa9ac02f2deddd03c6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:04:26.551616Z","signature_b64":"0OXWt562nW1zIENz0ASodFFolWqR/RTJ85NQETQ3fiRJspYA7AWQC4YASpaObp1leRO3BNdNY2OT99wT0UfHAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"bcb4060322511770452d04ea5859beb478e94824867fcbfba46ef2063b0fd5a9","last_reissued_at":"2026-07-05T10:04:26.551116Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:04:26.551116Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Watching the AI Watchdogs: A Fairness and Robustness Analysis of AI Safety Moderation Classifiers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Akshit Achara, Anshuman Chhabra","submitted_at":"2025-01-23T01:04:00Z","abstract_excerpt":"AI Safety Moderation (ASM) classifiers are designed to moderate content on social media platforms and to serve as guardrails that prevent Large Language Models (LLMs) from being fine-tuned on unsafe inputs. Owing to their potential for disparate impact, it is crucial to ensure that these classifiers: (1) do not unfairly classify content belonging to users from minority groups as unsafe compared to those from majority groups and (2) that their behavior remains robust and consistent across similar inputs. In this work, we thus examine the fairness and robustness of four widely-used, closed-sourc"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.13302","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/2501.13302/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":"2501.13302","created_at":"2026-07-05T10:04:26.551169+00:00"},{"alias_kind":"arxiv_version","alias_value":"2501.13302v1","created_at":"2026-07-05T10:04:26.551169+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.13302","created_at":"2026-07-05T10:04:26.551169+00:00"},{"alias_kind":"pith_short_12","alias_value":"XS2AMAZCKELX","created_at":"2026-07-05T10:04:26.551169+00:00"},{"alias_kind":"pith_short_16","alias_value":"XS2AMAZCKELXARJN","created_at":"2026-07-05T10:04:26.551169+00:00"},{"alias_kind":"pith_short_8","alias_value":"XS2AMAZC","created_at":"2026-07-05T10:04:26.551169+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.06552","citing_title":"To Lie or Not to Lie? Investigating The Biased Spread of Global Lies by LLMs","ref_index":4,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/XS2AMAZCKELXARJNATVFQWN6WR","json":"https://pith.science/pith/XS2AMAZCKELXARJNATVFQWN6WR.json","graph_json":"https://pith.science/api/pith-number/XS2AMAZCKELXARJNATVFQWN6WR/graph.json","events_json":"https://pith.science/api/pith-number/XS2AMAZCKELXARJNATVFQWN6WR/events.json","paper":"https://pith.science/paper/XS2AMAZC"},"agent_actions":{"view_html":"https://pith.science/pith/XS2AMAZCKELXARJNATVFQWN6WR","download_json":"https://pith.science/pith/XS2AMAZCKELXARJNATVFQWN6WR.json","view_paper":"https://pith.science/paper/XS2AMAZC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2501.13302&json=true","fetch_graph":"https://pith.science/api/pith-number/XS2AMAZCKELXARJNATVFQWN6WR/graph.json","fetch_events":"https://pith.science/api/pith-number/XS2AMAZCKELXARJNATVFQWN6WR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/XS2AMAZCKELXARJNATVFQWN6WR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/XS2AMAZCKELXARJNATVFQWN6WR/action/storage_attestation","attest_author":"https://pith.science/pith/XS2AMAZCKELXARJNATVFQWN6WR/action/author_attestation","sign_citation":"https://pith.science/pith/XS2AMAZCKELXARJNATVFQWN6WR/action/citation_signature","submit_replication":"https://pith.science/pith/XS2AMAZCKELXARJNATVFQWN6WR/action/replication_record"}},"created_at":"2026-07-05T10:04:26.551169+00:00","updated_at":"2026-07-05T10:04:26.551169+00:00"}