{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:FRECMOC5YIOCC3NRH7KODM3REF","short_pith_number":"pith:FRECMOC5","schema_version":"1.0","canonical_sha256":"2c4826385dc21c216db13fd4e1b3712155ea3bf9e22408b028eadb96921afc09","source":{"kind":"arxiv","id":"2507.22358","version":1},"attestation_state":"computed","paper":{"title":"Magentic-UI: Towards Human-in-the-loop Agentic Systems","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.HC"],"primary_cat":"cs.AI","authors_text":"Adam Fourney, Cheng Tan, Ece Kamar, Eric Zhu, Friederike Niedtner, Gagan Bansal, Griffin Bassman, Hussein Mozannar, Jack Gerrits, Jacob Alber, Jingya Chen, Madeleine Grunde-McLaughlin, Matheus Kunzler Maldaner, Maya Murad, Peter Chang, Rafah Hosn, Ricky Loynd, Saleema Amershi, Tyler Payne, Victor Dibia","submitted_at":"2025-07-30T03:49:14Z","abstract_excerpt":"AI agents powered by large language models are increasingly capable of autonomously completing complex, multi-step tasks using external tools. Yet, they still fall short of human-level performance in most domains including computer use, software development, and research. Their growing autonomy and ability to interact with the outside world, also introduces safety and security risks including potentially misaligned actions and adversarial manipulation. We argue that human-in-the-loop agentic systems offer a promising path forward, combining human oversight and control with AI efficiency to unl"},"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":"2507.22358","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2025-07-30T03:49:14Z","cross_cats_sorted":["cs.HC"],"title_canon_sha256":"2686e9335304242b027e05aa30c48a65ab0b2e46f2ef43845c002df18d7e9eca","abstract_canon_sha256":"a51e49d9af39e4dce78148a672560f5591588ce61284b66921268233519c110d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:45:40.076855Z","signature_b64":"EW3Y0Hz8sekoFw/7PqT6eNykcwpec4E+NTe827LDwiFMw9CZvZvc6Drros0QLF4Inol4iumWyQ1SJI7n0K+LBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2c4826385dc21c216db13fd4e1b3712155ea3bf9e22408b028eadb96921afc09","last_reissued_at":"2026-07-05T11:45:40.076353Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:45:40.076353Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Magentic-UI: Towards Human-in-the-loop Agentic Systems","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.HC"],"primary_cat":"cs.AI","authors_text":"Adam Fourney, Cheng Tan, Ece Kamar, Eric Zhu, Friederike Niedtner, Gagan Bansal, Griffin Bassman, Hussein Mozannar, Jack Gerrits, Jacob Alber, Jingya Chen, Madeleine Grunde-McLaughlin, Matheus Kunzler Maldaner, Maya Murad, Peter Chang, Rafah Hosn, Ricky Loynd, Saleema Amershi, Tyler Payne, Victor Dibia","submitted_at":"2025-07-30T03:49:14Z","abstract_excerpt":"AI agents powered by large language models are increasingly capable of autonomously completing complex, multi-step tasks using external tools. Yet, they still fall short of human-level performance in most domains including computer use, software development, and research. Their growing autonomy and ability to interact with the outside world, also introduces safety and security risks including potentially misaligned actions and adversarial manipulation. We argue that human-in-the-loop agentic systems offer a promising path forward, combining human oversight and control with AI efficiency to unl"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.22358","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/2507.22358/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":"2507.22358","created_at":"2026-07-05T11:45:40.076416+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.22358v1","created_at":"2026-07-05T11:45:40.076416+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.22358","created_at":"2026-07-05T11:45:40.076416+00:00"},{"alias_kind":"pith_short_12","alias_value":"FRECMOC5YIOC","created_at":"2026-07-05T11:45:40.076416+00:00"},{"alias_kind":"pith_short_16","alias_value":"FRECMOC5YIOCC3NR","created_at":"2026-07-05T11:45:40.076416+00:00"},{"alias_kind":"pith_short_8","alias_value":"FRECMOC5","created_at":"2026-07-05T11:45:40.076416+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":10,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.22484","citing_title":"Governed AI-Assisted Engineering: Graduated Human Oversight for Agentic Code Generation in Regulated Domains","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2606.18671","citing_title":"HANSEL: Extracting Breadcrumbs from Web Agent Trajectories for Interactive Verification","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2605.23023","citing_title":"How to Steer Your Multi-Agent System: Human-LLM Collaborative Planning","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2512.10371","citing_title":"AgentProg: Empowering Long-Horizon GUI Agents with Program-Guided Context Management","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11495","citing_title":"Hedwig: Dynamic Autonomy for Coding Agents Under Local Oversight","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2605.00798","citing_title":"RunAgent: Interpreting Natural-Language Plans with Constraint-Guided Execution","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2604.08898","citing_title":"Omakase: proactive assistance with actionable suggestions for evolving scientific research projects","ref_index":40,"is_internal_anchor":false},{"citing_arxiv_id":"2604.04918","citing_title":"Comparing Human Oversight Strategies for Computer-Use Agents","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2604.16520","citing_title":"AgentClick: A Skill-Based Human-in-the-Loop Review Layer for Terminal AI Agents","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2604.20070","citing_title":"Auditing and Controlling AI Agent Actions in Spreadsheets","ref_index":36,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FRECMOC5YIOCC3NRH7KODM3REF","json":"https://pith.science/pith/FRECMOC5YIOCC3NRH7KODM3REF.json","graph_json":"https://pith.science/api/pith-number/FRECMOC5YIOCC3NRH7KODM3REF/graph.json","events_json":"https://pith.science/api/pith-number/FRECMOC5YIOCC3NRH7KODM3REF/events.json","paper":"https://pith.science/paper/FRECMOC5"},"agent_actions":{"view_html":"https://pith.science/pith/FRECMOC5YIOCC3NRH7KODM3REF","download_json":"https://pith.science/pith/FRECMOC5YIOCC3NRH7KODM3REF.json","view_paper":"https://pith.science/paper/FRECMOC5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.22358&json=true","fetch_graph":"https://pith.science/api/pith-number/FRECMOC5YIOCC3NRH7KODM3REF/graph.json","fetch_events":"https://pith.science/api/pith-number/FRECMOC5YIOCC3NRH7KODM3REF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FRECMOC5YIOCC3NRH7KODM3REF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FRECMOC5YIOCC3NRH7KODM3REF/action/storage_attestation","attest_author":"https://pith.science/pith/FRECMOC5YIOCC3NRH7KODM3REF/action/author_attestation","sign_citation":"https://pith.science/pith/FRECMOC5YIOCC3NRH7KODM3REF/action/citation_signature","submit_replication":"https://pith.science/pith/FRECMOC5YIOCC3NRH7KODM3REF/action/replication_record"}},"created_at":"2026-07-05T11:45:40.076416+00:00","updated_at":"2026-07-05T11:45:40.076416+00:00"}