{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:ZSLLEW66STH55WDX7GPIZB236M","short_pith_number":"pith:ZSLLEW66","schema_version":"1.0","canonical_sha256":"cc96b25bde94cfded877f99e8c875bf30ea98c082dfa3588c82650b7b4725727","source":{"kind":"arxiv","id":"2305.06972","version":3},"attestation_state":"computed","paper":{"title":"Spear Phishing With Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CR"],"primary_cat":"cs.CY","authors_text":"Julian Hazell","submitted_at":"2023-05-11T16:55:19Z","abstract_excerpt":"Recent progress in artificial intelligence (AI), particularly in the domain of large language models (LLMs), has resulted in powerful and versatile dual-use systems. This intelligence can be put towards a wide variety of beneficial tasks, yet it can also be used to cause harm. This study explores one such harm by examining how LLMs can be used for spear phishing, a form of cybercrime that involves manipulating targets into divulging sensitive information. I first explore LLMs' ability to assist with the reconnaissance and message generation stages of a spear phishing attack, where I find that "},"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":"2305.06972","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CY","submitted_at":"2023-05-11T16:55:19Z","cross_cats_sorted":["cs.AI","cs.CR"],"title_canon_sha256":"8e0767aa6b570bec395ec9795dc403040a3862a9c16f1a5316774bc80f440585","abstract_canon_sha256":"d091501aa474c6eb52628efc2539b8f106b3a31fbccbc200148c146e0d8f7c26"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:27:04.444525Z","signature_b64":"mRb/O4Os+R/ptk8XeTNN1M13GHcrIipF9VpMIpQiy5y2AexBn90Tdzv+jm1+lDt/XbbfmHQQ+87m/dPF6iVHAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cc96b25bde94cfded877f99e8c875bf30ea98c082dfa3588c82650b7b4725727","last_reissued_at":"2026-07-05T07:27:04.444061Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:27:04.444061Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Spear Phishing With Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CR"],"primary_cat":"cs.CY","authors_text":"Julian Hazell","submitted_at":"2023-05-11T16:55:19Z","abstract_excerpt":"Recent progress in artificial intelligence (AI), particularly in the domain of large language models (LLMs), has resulted in powerful and versatile dual-use systems. This intelligence can be put towards a wide variety of beneficial tasks, yet it can also be used to cause harm. This study explores one such harm by examining how LLMs can be used for spear phishing, a form of cybercrime that involves manipulating targets into divulging sensitive information. I first explore LLMs' ability to assist with the reconnaissance and message generation stages of a spear phishing attack, where I find that "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.06972","kind":"arxiv","version":3},"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/2305.06972/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":"2305.06972","created_at":"2026-07-05T07:27:04.444121+00:00"},{"alias_kind":"arxiv_version","alias_value":"2305.06972v3","created_at":"2026-07-05T07:27:04.444121+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.06972","created_at":"2026-07-05T07:27:04.444121+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZSLLEW66STH5","created_at":"2026-07-05T07:27:04.444121+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZSLLEW66STH55WDX","created_at":"2026-07-05T07:27:04.444121+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZSLLEW66","created_at":"2026-07-05T07:27:04.444121+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":18,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.02423","citing_title":"Investigating and Alleviating Harm Amplification in LLM Interactions","ref_index":25,"is_internal_anchor":false},{"citing_arxiv_id":"2606.00160","citing_title":"DataShield: Safety-degrading Data Filtering for LLM Benign Instruction Fine-Tuning","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10977","citing_title":"PASA: A Principled Embedding-Space Watermarking Approach for LLM-Generated Text under Semantic-Invariant Attacks","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2605.16436","citing_title":"The End of Trust: How Agentic AI Breaks Security Assumptions","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2605.16471","citing_title":"From AI-Generated Content to Agentic Action: Security and Safety Threats in Generative AI","ref_index":53,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18239","citing_title":"Multilingual jailbreaking of LLMs using low-resource languages","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2508.21457","citing_title":"SoK: Exposing the Generation and Detection Gaps in LLM-Generated Phishing","ref_index":45,"is_internal_anchor":false},{"citing_arxiv_id":"2401.05561","citing_title":"TrustLLM: Trustworthiness in Large Language Models","ref_index":248,"is_internal_anchor":false},{"citing_arxiv_id":"2308.03825","citing_title":"\"Do Anything Now\": Characterizing and Evaluating In-The-Wild Jailbreak Prompts on Large Language Models","ref_index":34,"is_internal_anchor":false},{"citing_arxiv_id":"2310.06987","citing_title":"Catastrophic Jailbreak of Open-source LLMs via Exploiting Generation","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2307.02483","citing_title":"Jailbroken: How Does LLM Safety Training Fail?","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2604.03121","citing_title":"An Independent Safety Evaluation of Kimi K2.5","ref_index":79,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10977","citing_title":"PASA: A Principled Embedding-Space Watermarking Approach for LLM-Generated Text under Semantic-Invariant Attacks","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2310.04451","citing_title":"AutoDAN: Generating Stealthy Jailbreak Prompts on Aligned Large Language Models","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06524","citing_title":"Process Matters more than Output for Distinguishing Humans from Machines","ref_index":63,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06524","citing_title":"Process Matters more than Output for Distinguishing Humans from Machines","ref_index":63,"is_internal_anchor":false},{"citing_arxiv_id":"2604.10893","citing_title":"Beyond A Fixed Seal: Adaptive Stealing Watermark in Large Language Models","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2604.09544","citing_title":"Large Language Models Generate Harmful Responses Using a Distinct Mechanism, Shared Across Harm Types","ref_index":12,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ZSLLEW66STH55WDX7GPIZB236M","json":"https://pith.science/pith/ZSLLEW66STH55WDX7GPIZB236M.json","graph_json":"https://pith.science/api/pith-number/ZSLLEW66STH55WDX7GPIZB236M/graph.json","events_json":"https://pith.science/api/pith-number/ZSLLEW66STH55WDX7GPIZB236M/events.json","paper":"https://pith.science/paper/ZSLLEW66"},"agent_actions":{"view_html":"https://pith.science/pith/ZSLLEW66STH55WDX7GPIZB236M","download_json":"https://pith.science/pith/ZSLLEW66STH55WDX7GPIZB236M.json","view_paper":"https://pith.science/paper/ZSLLEW66","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2305.06972&json=true","fetch_graph":"https://pith.science/api/pith-number/ZSLLEW66STH55WDX7GPIZB236M/graph.json","fetch_events":"https://pith.science/api/pith-number/ZSLLEW66STH55WDX7GPIZB236M/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZSLLEW66STH55WDX7GPIZB236M/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZSLLEW66STH55WDX7GPIZB236M/action/storage_attestation","attest_author":"https://pith.science/pith/ZSLLEW66STH55WDX7GPIZB236M/action/author_attestation","sign_citation":"https://pith.science/pith/ZSLLEW66STH55WDX7GPIZB236M/action/citation_signature","submit_replication":"https://pith.science/pith/ZSLLEW66STH55WDX7GPIZB236M/action/replication_record"}},"created_at":"2026-07-05T07:27:04.444121+00:00","updated_at":"2026-07-05T07:27:04.444121+00:00"}