{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:CBSO3EW355TNAWORRAMESVCMRI","short_pith_number":"pith:CBSO3EW3","schema_version":"1.0","canonical_sha256":"1064ed92dbef66d059d1881849544c8a1cb73d8a8c1e9bd5672c267d91dbdf58","source":{"kind":"arxiv","id":"2603.09781","version":2},"attestation_state":"computed","paper":{"title":"CLIOPATRA: Extracting Private Information from LLM Insights","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CR","authors_text":"Emiliano De Cristofaro, Meenatchi Sundaram Muthu Selva Annamalai, Peter Kairouz","submitted_at":"2026-03-10T15:17:14Z","abstract_excerpt":"The widespread adoption of AI assistants has prompted the development of privacy-aware platforms designed to extract insights from real-world usage. Their privacy protections primarily rely on layering multiple heuristic techniques, such as PII redaction, clustering, aggregation, and LLM-based privacy auditing. In this paper, we put their privacy claims to the test by presenting CLIOPATRA, the first attack against ``privacy-preserving'' LLM-based insights systems. Our attack involves an adversary that carefully designs and inserts malicious chats into the system to break multiple layers of pro"},"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":"2603.09781","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2026-03-10T15:17:14Z","cross_cats_sorted":[],"title_canon_sha256":"da97226b1e98de6fa87f781108ff553d9eb9c50adc69291e31078680ea76d093","abstract_canon_sha256":"225f9defcce15608c6154f83a0e5566b13b651af4944ce94be434da5bf650a83"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-28T01:05:12.002402Z","signature_b64":"CmviCr9lu9MdLLuGKHtbuhW3jgfR/fEC6aXmYbE+tsyn/vxvzjVvdJmqjg+13r24myZ5i6XjVmXEi2gO8H3IDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1064ed92dbef66d059d1881849544c8a1cb73d8a8c1e9bd5672c267d91dbdf58","last_reissued_at":"2026-05-28T01:05:12.001919Z","signature_status":"signed_v1","first_computed_at":"2026-05-28T01:05:12.001919Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"CLIOPATRA: Extracting Private Information from LLM Insights","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CR","authors_text":"Emiliano De Cristofaro, Meenatchi Sundaram Muthu Selva Annamalai, Peter Kairouz","submitted_at":"2026-03-10T15:17:14Z","abstract_excerpt":"The widespread adoption of AI assistants has prompted the development of privacy-aware platforms designed to extract insights from real-world usage. Their privacy protections primarily rely on layering multiple heuristic techniques, such as PII redaction, clustering, aggregation, and LLM-based privacy auditing. In this paper, we put their privacy claims to the test by presenting CLIOPATRA, the first attack against ``privacy-preserving'' LLM-based insights systems. Our attack involves an adversary that carefully designs and inserts malicious chats into the system to break multiple layers of pro"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2603.09781","kind":"arxiv","version":2},"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/2603.09781/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":"2603.09781","created_at":"2026-05-28T01:05:12.001971+00:00"},{"alias_kind":"arxiv_version","alias_value":"2603.09781v2","created_at":"2026-05-28T01:05:12.001971+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2603.09781","created_at":"2026-05-28T01:05:12.001971+00:00"},{"alias_kind":"pith_short_12","alias_value":"CBSO3EW355TN","created_at":"2026-05-28T01:05:12.001971+00:00"},{"alias_kind":"pith_short_16","alias_value":"CBSO3EW355TNAWOR","created_at":"2026-05-28T01:05:12.001971+00:00"},{"alias_kind":"pith_short_8","alias_value":"CBSO3EW3","created_at":"2026-05-28T01:05:12.001971+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2605.12264","citing_title":"Reconstruction of Personally Identifiable Information from Supervised Finetuned Models","ref_index":3,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CBSO3EW355TNAWORRAMESVCMRI","json":"https://pith.science/pith/CBSO3EW355TNAWORRAMESVCMRI.json","graph_json":"https://pith.science/api/pith-number/CBSO3EW355TNAWORRAMESVCMRI/graph.json","events_json":"https://pith.science/api/pith-number/CBSO3EW355TNAWORRAMESVCMRI/events.json","paper":"https://pith.science/paper/CBSO3EW3"},"agent_actions":{"view_html":"https://pith.science/pith/CBSO3EW355TNAWORRAMESVCMRI","download_json":"https://pith.science/pith/CBSO3EW355TNAWORRAMESVCMRI.json","view_paper":"https://pith.science/paper/CBSO3EW3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2603.09781&json=true","fetch_graph":"https://pith.science/api/pith-number/CBSO3EW355TNAWORRAMESVCMRI/graph.json","fetch_events":"https://pith.science/api/pith-number/CBSO3EW355TNAWORRAMESVCMRI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CBSO3EW355TNAWORRAMESVCMRI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CBSO3EW355TNAWORRAMESVCMRI/action/storage_attestation","attest_author":"https://pith.science/pith/CBSO3EW355TNAWORRAMESVCMRI/action/author_attestation","sign_citation":"https://pith.science/pith/CBSO3EW355TNAWORRAMESVCMRI/action/citation_signature","submit_replication":"https://pith.science/pith/CBSO3EW355TNAWORRAMESVCMRI/action/replication_record"}},"created_at":"2026-05-28T01:05:12.001971+00:00","updated_at":"2026-05-28T01:05:12.001971+00:00"}