{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:BFKEDB66Y6JUSZXGZHNUWSXCDQ","short_pith_number":"pith:BFKEDB66","schema_version":"1.0","canonical_sha256":"09544187dec7934966e6c9db4b4ae21c26a026f5438e19eb8085c4ae88b19057","source":{"kind":"arxiv","id":"2504.10430","version":1},"attestation_state":"computed","paper":{"title":"LLM Can be a Dangerous Persuader: Empirical Study of Persuasion Safety in Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.HC"],"primary_cat":"cs.CL","authors_text":"Heajun An, Jin-Hee Cho, Lifu Huang, Minqian Liu, Pamela J. Wisniewski, Qi Zhang, Ruoxi Jia, Sang Won Lee, Sarvech Qadir, Xinyi Zhang, Zhiyang Xu","submitted_at":"2025-04-14T17:20:34Z","abstract_excerpt":"Recent advancements in Large Language Models (LLMs) have enabled them to approach human-level persuasion capabilities. However, such potential also raises concerns about the safety risks of LLM-driven persuasion, particularly their potential for unethical influence through manipulation, deception, exploitation of vulnerabilities, and many other harmful tactics. In this work, we present a systematic investigation of LLM persuasion safety through two critical aspects: (1) whether LLMs appropriately reject unethical persuasion tasks and avoid unethical strategies during execution, including cases"},"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":"2504.10430","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-04-14T17:20:34Z","cross_cats_sorted":["cs.AI","cs.HC"],"title_canon_sha256":"c73b506d84a28ad23882d4ef762cce1c9965963e4f6a72228f67ec0d0c11aa69","abstract_canon_sha256":"7b33e249c31cfb7593b75143f0d76737af0870dbb031a1f3730b11803b60e6b8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:49:00.226575Z","signature_b64":"dwmOlnBpe78KgwwcrmcCXRH2BIGl7sp2W3ZkkdH7nVf/bX1QY1Iw6dry5pU9RzrMazZUptY7eMfCxW3+YdQEBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"09544187dec7934966e6c9db4b4ae21c26a026f5438e19eb8085c4ae88b19057","last_reissued_at":"2026-07-05T10:49:00.226115Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:49:00.226115Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LLM Can be a Dangerous Persuader: Empirical Study of Persuasion Safety in Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.HC"],"primary_cat":"cs.CL","authors_text":"Heajun An, Jin-Hee Cho, Lifu Huang, Minqian Liu, Pamela J. Wisniewski, Qi Zhang, Ruoxi Jia, Sang Won Lee, Sarvech Qadir, Xinyi Zhang, Zhiyang Xu","submitted_at":"2025-04-14T17:20:34Z","abstract_excerpt":"Recent advancements in Large Language Models (LLMs) have enabled them to approach human-level persuasion capabilities. However, such potential also raises concerns about the safety risks of LLM-driven persuasion, particularly their potential for unethical influence through manipulation, deception, exploitation of vulnerabilities, and many other harmful tactics. In this work, we present a systematic investigation of LLM persuasion safety through two critical aspects: (1) whether LLMs appropriately reject unethical persuasion tasks and avoid unethical strategies during execution, including cases"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.10430","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/2504.10430/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":"2504.10430","created_at":"2026-07-05T10:49:00.226165+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.10430v1","created_at":"2026-07-05T10:49:00.226165+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.10430","created_at":"2026-07-05T10:49:00.226165+00:00"},{"alias_kind":"pith_short_12","alias_value":"BFKEDB66Y6JU","created_at":"2026-07-05T10:49:00.226165+00:00"},{"alias_kind":"pith_short_16","alias_value":"BFKEDB66Y6JUSZXG","created_at":"2026-07-05T10:49:00.226165+00:00"},{"alias_kind":"pith_short_8","alias_value":"BFKEDB66","created_at":"2026-07-05T10:49:00.226165+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.10126","citing_title":"Pareto-Guided Teacher Alignment for Fair Personalized Text Generation","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2606.06099","citing_title":"CogManip: Benchmarking Manipulative Behavior in Multi-Turn Interactions with Large Language Model","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2606.05330","citing_title":"A Model of Multi-turn Human Persuadability Using Probabilistic Belief Tracing","ref_index":75,"is_internal_anchor":false},{"citing_arxiv_id":"2510.12826","citing_title":"Scheming Ability in LLM-to-LLM Strategic Interactions","ref_index":36,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12772","citing_title":"Just Ask for a Table: A Thirty-Token User Prompt Defeats Sponsored Recommendations in Twelve LLMs","ref_index":6,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BFKEDB66Y6JUSZXGZHNUWSXCDQ","json":"https://pith.science/pith/BFKEDB66Y6JUSZXGZHNUWSXCDQ.json","graph_json":"https://pith.science/api/pith-number/BFKEDB66Y6JUSZXGZHNUWSXCDQ/graph.json","events_json":"https://pith.science/api/pith-number/BFKEDB66Y6JUSZXGZHNUWSXCDQ/events.json","paper":"https://pith.science/paper/BFKEDB66"},"agent_actions":{"view_html":"https://pith.science/pith/BFKEDB66Y6JUSZXGZHNUWSXCDQ","download_json":"https://pith.science/pith/BFKEDB66Y6JUSZXGZHNUWSXCDQ.json","view_paper":"https://pith.science/paper/BFKEDB66","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.10430&json=true","fetch_graph":"https://pith.science/api/pith-number/BFKEDB66Y6JUSZXGZHNUWSXCDQ/graph.json","fetch_events":"https://pith.science/api/pith-number/BFKEDB66Y6JUSZXGZHNUWSXCDQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BFKEDB66Y6JUSZXGZHNUWSXCDQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BFKEDB66Y6JUSZXGZHNUWSXCDQ/action/storage_attestation","attest_author":"https://pith.science/pith/BFKEDB66Y6JUSZXGZHNUWSXCDQ/action/author_attestation","sign_citation":"https://pith.science/pith/BFKEDB66Y6JUSZXGZHNUWSXCDQ/action/citation_signature","submit_replication":"https://pith.science/pith/BFKEDB66Y6JUSZXGZHNUWSXCDQ/action/replication_record"}},"created_at":"2026-07-05T10:49:00.226165+00:00","updated_at":"2026-07-05T10:49:00.226165+00:00"}