{"paper":{"title":"FieldWorkArena: Agentic AI Benchmark for Real Field Work Tasks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"FieldWorkArena uses real factory and retail photos to test whether agentic AI can spot safety hazards and rule violations on site.","cross_cats":["cs.CV"],"primary_cat":"cs.AI","authors_text":"Akiyoshi Uchida, Atsunori Moteki, Fan Yang, Graham Neubig, Hiroyuki Ishida, Ikuo Kusajima, Jun Takahashi, Kanji Uchino, Koki Nakagawa, Shan Jiang, Shoichi Masui, Yasuto Watanabe, Yonatan Bisk, Yueqi Song","submitted_at":"2025-05-26T08:21:46Z","abstract_excerpt":"This paper introduces FieldWorkArena, a benchmark for agentic AI targeting real-world field work. With the recent increase in demand for agentic AI, they are built to detect and document safety hazards, procedural violations, and other critical incidents across real-world manufacturing and retail environments. Whereas most agentic AI benchmarks focus on performance in simulated or digital environments, our work addresses the fundamental challenge of evaluating agents in the real-world. In this paper, we improve the evaluation function from previous methods to assess the performance of agentic "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Evaluation results confirmed that performance evaluation considering the characteristics of Multimodal LLM (MLLM) such as GPT-4o is feasible.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The assumption that on-site captured images/videos from factories, warehouses and retails combined with tasks developed through interviews with site workers and managers provide a representative and sufficient basis for evaluating agentic AI performance in real-world conditions.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"A new benchmark dataset and evaluation framework for testing multimodal AI agents on real field work tasks derived from on-site data and worker interviews.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"FieldWorkArena uses real factory and retail photos to test whether agentic AI can spot safety hazards and rule violations on site.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"deac717ef9882310340abf5bfe6c65388b6385bf56fe38f6d3256015f9a6e130"},"source":{"id":"2505.19662","kind":"arxiv","version":4},"verdict":{"id":"637eb298-c243-42f2-84d0-1c504a78d1de","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T14:32:07.079733Z","strongest_claim":"Evaluation results confirmed that performance evaluation considering the characteristics of Multimodal LLM (MLLM) such as GPT-4o is feasible.","one_line_summary":"A new benchmark dataset and evaluation framework for testing multimodal AI agents on real field work tasks derived from on-site data and worker interviews.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The assumption that on-site captured images/videos from factories, warehouses and retails combined with tasks developed through interviews with site workers and managers provide a representative and sufficient basis for evaluating agentic AI performance in real-world conditions.","pith_extraction_headline":"FieldWorkArena uses real factory and retail photos to test whether agentic AI can spot safety hazards and rule violations on site."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2505.19662/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"}