{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:MBWCOVAKNQWRW757A5VBHOC6JU","short_pith_number":"pith:MBWCOVAK","schema_version":"1.0","canonical_sha256":"606c27540a6c2d1b7fbf076a13b85e4d2989b1b13824c5670e2823dff5f5a046","source":{"kind":"arxiv","id":"2503.08663","version":1},"attestation_state":"computed","paper":{"title":"Generating Robot Constitutions & Benchmarks for Semantic Safety","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CV","cs.CY","cs.HC"],"primary_cat":"cs.RO","authors_text":"Alex Irpan, Anirudha Majumdar, Dmitry Kalashnikov, Pierre Sermanet, Vikas Sindhwani","submitted_at":"2025-03-11T17:50:47Z","abstract_excerpt":"Until recently, robotics safety research was predominantly about collision avoidance and hazard reduction in the immediate vicinity of a robot. Since the advent of large vision and language models (VLMs), robots are now also capable of higher-level semantic scene understanding and natural language interactions with humans. Despite their known vulnerabilities (e.g. hallucinations or jail-breaking), VLMs are being handed control of robots capable of physical contact with the real world. This can lead to dangerous behaviors, making semantic safety for robots a matter of immediate concern. Our con"},"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":"2503.08663","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.RO","submitted_at":"2025-03-11T17:50:47Z","cross_cats_sorted":["cs.AI","cs.CV","cs.CY","cs.HC"],"title_canon_sha256":"f3de13c1d7a037ba8bbdc48e02325af9a693c1033c1449d7a707105fa40bbc21","abstract_canon_sha256":"ddcab58c910c6e2e82c6a7eec19b0e050a4bfe814517314b5ef1ac9c596b5697"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:29:04.613313Z","signature_b64":"Wpm8mZKMsMFMJVedHwMFOwt8CwxkdHtlpQs2npYZKd1w1mwQd/3cq6f3IUaqrCr3erYgENGmPWNXKuv/+WjSDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"606c27540a6c2d1b7fbf076a13b85e4d2989b1b13824c5670e2823dff5f5a046","last_reissued_at":"2026-07-05T10:29:04.612692Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:29:04.612692Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Generating Robot Constitutions & Benchmarks for Semantic Safety","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CV","cs.CY","cs.HC"],"primary_cat":"cs.RO","authors_text":"Alex Irpan, Anirudha Majumdar, Dmitry Kalashnikov, Pierre Sermanet, Vikas Sindhwani","submitted_at":"2025-03-11T17:50:47Z","abstract_excerpt":"Until recently, robotics safety research was predominantly about collision avoidance and hazard reduction in the immediate vicinity of a robot. Since the advent of large vision and language models (VLMs), robots are now also capable of higher-level semantic scene understanding and natural language interactions with humans. Despite their known vulnerabilities (e.g. hallucinations or jail-breaking), VLMs are being handed control of robots capable of physical contact with the real world. This can lead to dangerous behaviors, making semantic safety for robots a matter of immediate concern. Our con"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.08663","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/2503.08663/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":"2503.08663","created_at":"2026-07-05T10:29:04.612768+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.08663v1","created_at":"2026-07-05T10:29:04.612768+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.08663","created_at":"2026-07-05T10:29:04.612768+00:00"},{"alias_kind":"pith_short_12","alias_value":"MBWCOVAKNQWR","created_at":"2026-07-05T10:29:04.612768+00:00"},{"alias_kind":"pith_short_16","alias_value":"MBWCOVAKNQWRW757","created_at":"2026-07-05T10:29:04.612768+00:00"},{"alias_kind":"pith_short_8","alias_value":"MBWCOVAK","created_at":"2026-07-05T10:29:04.612768+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":14,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.27079","citing_title":"ForesightSafety-VLA: A Unified Diagnostic Safety Benchmark for Vision-Language-Action Models","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2606.17511","citing_title":"MagicSim: A Unified Infrastructure for Executable Embodied Interaction","ref_index":91,"is_internal_anchor":false},{"citing_arxiv_id":"2607.00218","citing_title":"EgoSafetyBench: A Diagnostic Egocentric Video Benchmark for Evaluating Embodied VLMs as Runtime Safety Guards","ref_index":25,"is_internal_anchor":false},{"citing_arxiv_id":"2606.03954","citing_title":"VLESA: Vision-Language Embodied Safety Agent for Human Activity Monitoring","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2606.27079","citing_title":"ForesightSafety-VLA: A Unified Diagnostic Safety Benchmark for Vision-Language-Action Models","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2606.01036","citing_title":"Position: Good Embodied Reward Models Need Bad Behavior Data","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2510.13727","citing_title":"From Refusal to Recovery: A Control-Theoretic Approach to Generative AI Guardrails","ref_index":80,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18423","citing_title":"REBAR: Reference Ethical Benchmark for Autonomy Readiness","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2506.05171","citing_title":"Towards provable probabilistic safety for scalable embodied AI systems","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2509.01728","citing_title":"Constrained Decoding for Safe Robot Navigation Foundation Models","ref_index":37,"is_internal_anchor":false},{"citing_arxiv_id":"2604.23775","citing_title":"Vision-Language-Action Safety: Threats, Challenges, Evaluations, and Mechanisms","ref_index":60,"is_internal_anchor":false},{"citing_arxiv_id":"2604.19638","citing_title":"SafetyALFRED: Evaluating Safety-Conscious Planning of Multimodal Large Language Models","ref_index":35,"is_internal_anchor":false},{"citing_arxiv_id":"2604.18463","citing_title":"Using large language models for embodied planning introduces systematic safety risks","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2604.19638","citing_title":"SafetyALFRED: Evaluating Safety-Conscious Planning of Multimodal Large Language Models","ref_index":21,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MBWCOVAKNQWRW757A5VBHOC6JU","json":"https://pith.science/pith/MBWCOVAKNQWRW757A5VBHOC6JU.json","graph_json":"https://pith.science/api/pith-number/MBWCOVAKNQWRW757A5VBHOC6JU/graph.json","events_json":"https://pith.science/api/pith-number/MBWCOVAKNQWRW757A5VBHOC6JU/events.json","paper":"https://pith.science/paper/MBWCOVAK"},"agent_actions":{"view_html":"https://pith.science/pith/MBWCOVAKNQWRW757A5VBHOC6JU","download_json":"https://pith.science/pith/MBWCOVAKNQWRW757A5VBHOC6JU.json","view_paper":"https://pith.science/paper/MBWCOVAK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.08663&json=true","fetch_graph":"https://pith.science/api/pith-number/MBWCOVAKNQWRW757A5VBHOC6JU/graph.json","fetch_events":"https://pith.science/api/pith-number/MBWCOVAKNQWRW757A5VBHOC6JU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MBWCOVAKNQWRW757A5VBHOC6JU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MBWCOVAKNQWRW757A5VBHOC6JU/action/storage_attestation","attest_author":"https://pith.science/pith/MBWCOVAKNQWRW757A5VBHOC6JU/action/author_attestation","sign_citation":"https://pith.science/pith/MBWCOVAKNQWRW757A5VBHOC6JU/action/citation_signature","submit_replication":"https://pith.science/pith/MBWCOVAKNQWRW757A5VBHOC6JU/action/replication_record"}},"created_at":"2026-07-05T10:29:04.612768+00:00","updated_at":"2026-07-05T10:29:04.612768+00:00"}