{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:H3HEBSQWDCB5Z6ZH2EFJ7WT2OJ","short_pith_number":"pith:H3HEBSQW","schema_version":"1.0","canonical_sha256":"3ece40ca161883dcfb27d10a9fda7a727b1104751c47a6a40428dfb5e94c2b41","source":{"kind":"arxiv","id":"2605.14218","version":1},"attestation_state":"computed","paper":{"title":"Fusion-fission forecasts when AI will shift to undesirable behavior","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"A vector generalization of fusion-fission group dynamics forecasts when AI behavior shifts from desirable to undesirable.","cross_cats":["physics.soc-ph"],"primary_cat":"cs.AI","authors_text":"Frank Yingjie Huo, Neil F. Johnson","submitted_at":"2026-05-14T00:26:32Z","abstract_excerpt":"The key problem facing ChatGPT-like AI's use across society is that its behavior can shift, unnoticed, from desirable to undesirable -- encouraging self-harm, extremist acts, financial losses, or costly medical and military mistakes -- and no one can yet predict when. Shifts persist in even the newest AI models despite remarkable progress in AI modeling, post-training alignment and safeguards. Here we show that a vector generalization of fusion-fission group dynamics observed in living and active-matter systems drives -- and can forecast -- future shifts in the AI's behavior. The shift conditi"},"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":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2605.14218","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-05-14T00:26:32Z","cross_cats_sorted":["physics.soc-ph"],"title_canon_sha256":"b2804448c4365bcea2d08ddc8390dfe3ea8e0e23c6065f7af6d673e9e433a265","abstract_canon_sha256":"a8d7e31dd133152178e38700d8bc5f4b24af0b94b8bcf572873002bfa3f143e4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:39:10.855445Z","signature_b64":"BQ+sDMKSPOzDX1yda5Tu1RlESHl70V5MOXuAEcZGR2DdpHUpPFvx63HE45YRDP+jNw0QVPegJE2ive7Qt4GTAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3ece40ca161883dcfb27d10a9fda7a727b1104751c47a6a40428dfb5e94c2b41","last_reissued_at":"2026-05-17T23:39:10.854636Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:39:10.854636Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Fusion-fission forecasts when AI will shift to undesirable behavior","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"A vector generalization of fusion-fission group dynamics forecasts when AI behavior shifts from desirable to undesirable.","cross_cats":["physics.soc-ph"],"primary_cat":"cs.AI","authors_text":"Frank Yingjie Huo, Neil F. Johnson","submitted_at":"2026-05-14T00:26:32Z","abstract_excerpt":"The key problem facing ChatGPT-like AI's use across society is that its behavior can shift, unnoticed, from desirable to undesirable -- encouraging self-harm, extremist acts, financial losses, or costly medical and military mistakes -- and no one can yet predict when. Shifts persist in even the newest AI models despite remarkable progress in AI modeling, post-training alignment and safeguards. Here we show that a vector generalization of fusion-fission group dynamics observed in living and active-matter systems drives -- and can forecast -- future shifts in the AI's behavior. The shift conditi"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"a vector generalization of fusion-fission group dynamics observed in living and active-matter systems drives -- and can forecast -- future shifts in the AI's behavior. The shift condition... results from group-level competition between the conversation-so-far (C) and the desirable (B) and undesirable (D) basin dynamics which can be estimated in advance for a given application.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the desirable (B) and undesirable (D) basin dynamics can be estimated in advance for any given application and that the vector generalization of fusion-fission dynamics actually governs AI conversation trajectories rather than merely correlating with them after the fact.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"A vector generalization of fusion-fission group dynamics from physics forecasts when AI behavior shifts to undesirable states, validated at 90 percent across seven models and prior to real-world data.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A vector generalization of fusion-fission group dynamics forecasts when AI behavior shifts from desirable to undesirable.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"b023d5ada952a8fa9d5c68d0eb83370c0f80f7370c852e557c6fa2c4ef51a6d8"},"source":{"id":"2605.14218","kind":"arxiv","version":1},"verdict":{"id":"417a8423-7824-44aa-9264-67d88e6db387","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T02:37:43.840844Z","strongest_claim":"a vector generalization of fusion-fission group dynamics observed in living and active-matter systems drives -- and can forecast -- future shifts in the AI's behavior. The shift condition... results from group-level competition between the conversation-so-far (C) and the desirable (B) and undesirable (D) basin dynamics which can be estimated in advance for a given application.","one_line_summary":"A vector generalization of fusion-fission group dynamics from physics forecasts when AI behavior shifts to undesirable states, validated at 90 percent across seven models and prior to real-world data.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the desirable (B) and undesirable (D) basin dynamics can be estimated in advance for any given application and that the vector generalization of fusion-fission dynamics actually governs AI conversation trajectories rather than merely correlating with them after the fact.","pith_extraction_headline":"A vector generalization of fusion-fission group dynamics forecasts when AI behavior shifts from desirable to undesirable."},"references":{"count":57,"sample":[{"doi":"","year":2025,"title":"CCDH report, 6 August 2025.https://counterhate.com/re search/fake-friend-chatgpt/","work_id":"2aed709b-2779-470e-bb43-24e016a685dd","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2026,"title":"URL http://arxiv.org/abs/2603.16567","work_id":"ec68e169-9fe1-4435-aafd-4fd39a381d53","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2026,"title":"CCDH report, 11 March 2026.https://counterhate.com/research /killer-apps/","work_id":"e7faa678-5ff5-45bc-b9e5-d0e1e323363f","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2026,"title":"O’Donnell, The new war room.MIT Technology Review(21 April 2026).https://www.tech nologyreview.com/2026/04/21/1135667/new-war-room-military-ai-artificial-intel ligence/","work_id":"d5fd477b-f0ff-45d9-8f04-1e1b164d0bde","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2021,"title":"Elhage et al., A mathematical framework for transformer circuits.Transformer Circuits Thread(2021).https://transformer-circuits.pub/2021/framework/index.html","work_id":"938b6822-6775-4a54-a2d0-fa0a1611a502","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":57,"snapshot_sha256":"d1a04d2cf202dd2011159969712f0ac793f1a334a6053c82bbfb66aa63d6669b","internal_anchors":8},"formal_canon":{"evidence_count":2,"snapshot_sha256":"f330782264ef07f070ff9cec27d427c459b5473a58fca5caf28cbe1d88c90412"},"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":"2605.14218","created_at":"2026-05-17T23:39:10.855031+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.14218v1","created_at":"2026-05-17T23:39:10.855031+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.14218","created_at":"2026-05-17T23:39:10.855031+00:00"},{"alias_kind":"pith_short_12","alias_value":"H3HEBSQWDCB5","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"H3HEBSQWDCB5Z6ZH","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"H3HEBSQW","created_at":"2026-05-18T12:33:37.589309+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/H3HEBSQWDCB5Z6ZH2EFJ7WT2OJ","json":"https://pith.science/pith/H3HEBSQWDCB5Z6ZH2EFJ7WT2OJ.json","graph_json":"https://pith.science/api/pith-number/H3HEBSQWDCB5Z6ZH2EFJ7WT2OJ/graph.json","events_json":"https://pith.science/api/pith-number/H3HEBSQWDCB5Z6ZH2EFJ7WT2OJ/events.json","paper":"https://pith.science/paper/H3HEBSQW"},"agent_actions":{"view_html":"https://pith.science/pith/H3HEBSQWDCB5Z6ZH2EFJ7WT2OJ","download_json":"https://pith.science/pith/H3HEBSQWDCB5Z6ZH2EFJ7WT2OJ.json","view_paper":"https://pith.science/paper/H3HEBSQW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.14218&json=true","fetch_graph":"https://pith.science/api/pith-number/H3HEBSQWDCB5Z6ZH2EFJ7WT2OJ/graph.json","fetch_events":"https://pith.science/api/pith-number/H3HEBSQWDCB5Z6ZH2EFJ7WT2OJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/H3HEBSQWDCB5Z6ZH2EFJ7WT2OJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/H3HEBSQWDCB5Z6ZH2EFJ7WT2OJ/action/storage_attestation","attest_author":"https://pith.science/pith/H3HEBSQWDCB5Z6ZH2EFJ7WT2OJ/action/author_attestation","sign_citation":"https://pith.science/pith/H3HEBSQWDCB5Z6ZH2EFJ7WT2OJ/action/citation_signature","submit_replication":"https://pith.science/pith/H3HEBSQWDCB5Z6ZH2EFJ7WT2OJ/action/replication_record"}},"created_at":"2026-05-17T23:39:10.855031+00:00","updated_at":"2026-05-17T23:39:10.855031+00:00"}