{"paper":{"title":"Identifying Harm in Personalized, Generative AI Systems Requires User-Centered Auditing at the Interaction Level","arxiv_id":"2608.14692","primary_cat":"cs.CY","submitted_at":"2026-08-07T15:58:37+00:00","version_compiled":"v1"},"notice":"Every quoted sentence is the authors' text from the arXiv listing abstract. Labels and probabilities are the outputs of classifier-1.13.0 on the question bank abstract-p1-20260923, exactly as returned. No sentence in this record was written by a language model.","schema":"pith.compiled-record/v0","status":"compiled","abstract":{"source":"arXiv listing abstract v1","sentences":[{"i":0,"p":1.0,"id":"s0","own":false,"role":"background","text":"Personalized, generative AI systems increasingly adapt their behavior to individual users over time, fundamentally changing model behavior.","hedge":{"p":0.99,"label":"asserted"},"own_p":0.07,"main_claim_p":0.03,"quantitative":false,"quantitative_p":0.06,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.0,"background":1.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":1,"p":0.51,"id":"s1","own":false,"role":"background","text":"While existing auditing approaches have been effective at surfacing harms in non-personalized contexts, they often rely on static, simulated evaluations and definitions of harm that aggregate across broad, group categories.","hedge":{"p":0.76,"label":"asserted"},"own_p":0.04,"main_claim_p":0.03,"quantitative":false,"quantitative_p":0.27,"role_probabilities":{"gap":0.48,"method":0.0,"result":0.0,"background":0.51,"limitation":0.01,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":2,"p":0.85,"id":"s2","own":false,"role":"gap","text":"In this position paper, we argue that such approaches can fail to capture emergent harms in personalized generative AI systems, where harms surface through interpretations of ongoing interaction and evolve with user history.","hedge":{"p":0.59,"label":"hedged_possibility"},"own_p":0.14,"main_claim_p":0.25,"quantitative":false,"quantitative_p":0.11,"role_probabilities":{"gap":0.85,"method":0.0,"result":0.0,"background":0.03,"limitation":0.0,"future_work":0.0,"contribution":0.12,"interpretation":0.0}},{"i":3,"p":0.84,"id":"s3","own":true,"role":"contribution","text":"We identify three presuppositions underlying many harm auditing paradigms: that harms can be (1) specified outside real-world interaction, (2) defined non-pluralistically within groups, and (3) treated as static.","hedge":{"p":0.98,"label":"asserted"},"own_p":0.56,"main_claim_p":0.07,"quantitative":true,"quantitative_p":0.84,"role_probabilities":{"gap":0.12,"method":0.0,"result":0.01,"background":0.02,"limitation":0.0,"future_work":0.0,"contribution":0.84,"interpretation":0.01}},{"i":4,"p":0.76,"id":"s4","own":false,"role":"background","text":"One might argue that personalized systems could simply learn definitions of what constitutes harm to individual users through repeated interactions.","hedge":{"p":0.81,"label":"speculative"},"own_p":0.04,"main_claim_p":0.02,"quantitative":false,"quantitative_p":0.04,"role_probabilities":{"gap":0.13,"method":0.0,"result":0.0,"background":0.76,"limitation":0.03,"future_work":0.0,"contribution":0.01,"interpretation":0.07}},{"i":5,"p":0.66,"id":"s5","own":false,"role":"interpretation","text":"However, we argue that attempts to surface user harms through deeper personalization risk imposing asymmetric burdens of labor and privacy on marginalized users.","hedge":{"p":0.6,"label":"asserted"},"own_p":0.08,"main_claim_p":0.08,"quantitative":true,"quantitative_p":0.52,"role_probabilities":{"gap":0.08,"method":0.0,"result":0.05,"background":0.03,"limitation":0.14,"future_work":0.0,"contribution":0.04,"interpretation":0.66}},{"i":6,"p":0.84,"id":"s6","own":true,"role":"contribution","text":"Consequently, we propose reframing understandings of harm as adaptive, user- and community-centered processes, and outline design directions that shift auditing from retrospective evaluation toward infrastructures that support ongoing articulation of harm in interaction.","hedge":{"p":0.98,"label":"asserted"},"own_p":0.56,"main_claim_p":0.85,"quantitative":false,"quantitative_p":0.07,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.16,"contribution":0.84,"interpretation":0.0}},{"i":7,"p":0.54,"id":"s7","own":false,"role":"interpretation","text":"Our work highlights the need for auditing and design practices that better reflect the pluralistic and evolving nature of harm understanding in personalized generative AI systems.","hedge":{"p":0.9,"label":"asserted"},"own_p":0.45,"main_claim_p":0.27,"quantitative":false,"quantitative_p":0.05,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.01,"background":0.0,"limitation":0.0,"future_work":0.36,"contribution":0.09,"interpretation":0.54}}]},"compiled":{"main_claim":{"p":0.93,"text":"Consequently, we propose reframing understandings of harm as adaptive, user- and community-centered processes, and outline design directions that shift auditing from retrospective evaluation toward infrastructures that support ongoing articulation of harm in interaction.","location":"abstract s6","hedge_strength":{"p":0.98,"label":"asserted"}},"provenance":{"route":{"via":"TypeSafe API","model":"classifier-1.13.0","endpoint":"https://api.typesafe.ai/v1/systemone","provider":"TypeSafe"},"packing":"one request per paper","requests":1,"input_tokens":6364,"cost_usd_reported":0.0},"compiled_at":"2026-09-23","bank_version":"abstract-p1-20260923","model_version":"classifier-1.13.0","abstract_states":{"claims_first":0.05,"states_limitation":0.15,"human_participants":0.21,"controlled_comparison":0.04,"randomized_assignment":0.01,"states_quantitative_result":0.02,"compares_with_prior_results":0.59,"mentions_code_or_data_availability":0.02},"evidential_basis":{"p":1.0,"value":"argument","probabilities":{"argument":1.0,"literature":0.0,"proof_or_derivation":0.0,"analysis_of_existing_data":0.0,"observation_or_experiment":0.0,"simulation_or_computation":0.0}},"contribution_type":{"p":1.0,"value":"position or commentary","probabilities":{"review":0.0,"method_or_tool":0.0,"empirical_study":0.0,"simulation_study":0.0,"theoretical_result":0.0,"dataset_or_benchmark":0.0,"position_or_commentary":1.0}},"contribution_statement":{"p":0.99,"text":"Consequently, we propose reframing understandings of harm as adaptive, user- and community-centered processes, and outline design directions that shift auditing from retrospective evaluation toward infrastructures that support ongoing articulation of harm in interaction.","location":"abstract s6"}},"coverage":{"tier":"abstract","levels_run":["P abstract"],"not_examined":["full text","figures and tables","references and citing sentences","relations to other papers"],"sources_read":["arXiv listing abstract v1"],"levels_not_run":["S sections","T body sentences","C citations","R cross-paper relations"],"sentences_examined":8,"sentences_in_abstract":8},"measured_accuracy":{"url":"https://pith.science/accuracy","measured_on":"2026-10-02","method":"two independent model families labelled a fresh random sample blind; the figure is how often the published label matches their consensus","sentence_role_agreement":0.92,"main_claim_agreement":0.908,"typed_citation_agreement":0.933,"typed_citation_agreement_95_interval":[0.917,0.945],"review_comments_accurate":0.758,"review_comments_accurate_95_interval":[0.684,0.82],"review_caught_admitted_error_found_or_related":0.448}}