{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:MVA64KQQWQWXAFPSKJ4Y44HAZ4","short_pith_number":"pith:MVA64KQQ","schema_version":"1.0","canonical_sha256":"6541ee2a10b42d7015f252798e70e0cf0da5c94e8dac8e45a982071236bc1b7a","source":{"kind":"arxiv","id":"2506.13609","version":1},"attestation_state":"computed","paper":{"title":"Avoiding Obfuscation with Prover-Estimator Debate","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CC","cs.DS"],"primary_cat":"cs.AI","authors_text":"Geoffrey Irving, Georgios Piliouras, Jonah Brown-Cohen","submitted_at":"2025-06-16T15:37:33Z","abstract_excerpt":"Training powerful AI systems to exhibit desired behaviors hinges on the ability to provide accurate human supervision on increasingly complex tasks. A promising approach to this problem is to amplify human judgement by leveraging the power of two competing AIs in a debate about the correct solution to a given problem. Prior theoretical work has provided a complexity-theoretic formalization of AI debate, and posed the problem of designing protocols for AI debate that guarantee the correctness of human judgements for as complex a class of problems as possible. Recursive debates, in which debater"},"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":"2506.13609","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2025-06-16T15:37:33Z","cross_cats_sorted":["cs.CC","cs.DS"],"title_canon_sha256":"e649783025d5bf5818725942e73d7bd29d4cd5e6e9a37422478304f0019318fa","abstract_canon_sha256":"9ad7873acc16e5595765641675f98cf28d040a795ac2fcc32c92a2cd12a9449a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:22:23.375370Z","signature_b64":"Ib75a/4pKKtKi6iLvrrkD22qUOE3WOtEAktqzMZaYNnxQJgkw7pB2c1b36wRJpQ5WAXiiD9lIziWzhfrfZpKAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6541ee2a10b42d7015f252798e70e0cf0da5c94e8dac8e45a982071236bc1b7a","last_reissued_at":"2026-07-05T11:22:23.374678Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:22:23.374678Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Avoiding Obfuscation with Prover-Estimator Debate","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CC","cs.DS"],"primary_cat":"cs.AI","authors_text":"Geoffrey Irving, Georgios Piliouras, Jonah Brown-Cohen","submitted_at":"2025-06-16T15:37:33Z","abstract_excerpt":"Training powerful AI systems to exhibit desired behaviors hinges on the ability to provide accurate human supervision on increasingly complex tasks. A promising approach to this problem is to amplify human judgement by leveraging the power of two competing AIs in a debate about the correct solution to a given problem. Prior theoretical work has provided a complexity-theoretic formalization of AI debate, and posed the problem of designing protocols for AI debate that guarantee the correctness of human judgements for as complex a class of problems as possible. Recursive debates, in which debater"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.13609","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/2506.13609/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":"2506.13609","created_at":"2026-07-05T11:22:23.374795+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.13609v1","created_at":"2026-07-05T11:22:23.374795+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.13609","created_at":"2026-07-05T11:22:23.374795+00:00"},{"alias_kind":"pith_short_12","alias_value":"MVA64KQQWQWX","created_at":"2026-07-05T11:22:23.374795+00:00"},{"alias_kind":"pith_short_16","alias_value":"MVA64KQQWQWXAFPS","created_at":"2026-07-05T11:22:23.374795+00:00"},{"alias_kind":"pith_short_8","alias_value":"MVA64KQQ","created_at":"2026-07-05T11:22:23.374795+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2605.28807","citing_title":"Calibrating Conservatism for Scalable Oversight","ref_index":2,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MVA64KQQWQWXAFPSKJ4Y44HAZ4","json":"https://pith.science/pith/MVA64KQQWQWXAFPSKJ4Y44HAZ4.json","graph_json":"https://pith.science/api/pith-number/MVA64KQQWQWXAFPSKJ4Y44HAZ4/graph.json","events_json":"https://pith.science/api/pith-number/MVA64KQQWQWXAFPSKJ4Y44HAZ4/events.json","paper":"https://pith.science/paper/MVA64KQQ"},"agent_actions":{"view_html":"https://pith.science/pith/MVA64KQQWQWXAFPSKJ4Y44HAZ4","download_json":"https://pith.science/pith/MVA64KQQWQWXAFPSKJ4Y44HAZ4.json","view_paper":"https://pith.science/paper/MVA64KQQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.13609&json=true","fetch_graph":"https://pith.science/api/pith-number/MVA64KQQWQWXAFPSKJ4Y44HAZ4/graph.json","fetch_events":"https://pith.science/api/pith-number/MVA64KQQWQWXAFPSKJ4Y44HAZ4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MVA64KQQWQWXAFPSKJ4Y44HAZ4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MVA64KQQWQWXAFPSKJ4Y44HAZ4/action/storage_attestation","attest_author":"https://pith.science/pith/MVA64KQQWQWXAFPSKJ4Y44HAZ4/action/author_attestation","sign_citation":"https://pith.science/pith/MVA64KQQWQWXAFPSKJ4Y44HAZ4/action/citation_signature","submit_replication":"https://pith.science/pith/MVA64KQQWQWXAFPSKJ4Y44HAZ4/action/replication_record"}},"created_at":"2026-07-05T11:22:23.374795+00:00","updated_at":"2026-07-05T11:22:23.374795+00:00"}