{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,7]],"date-time":"2026-09-07T14:26:38Z","timestamp":1788791198730,"version":"build-2803163510"},"reference-count":46,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,7,2]],"date-time":"2026-07-02T00:00:00Z","timestamp":1782950400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Artif. Intell."],"abstract":"<jats:sec>\n                    <jats:title>Introduction<\/jats:title>\n                    <jats:p>Autonomous and semi-automated AI tools are increasingly involved in digital-media workflows, creating a need for reliable preliminary screening of legal risks before publication or moderation decisions are made.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>This study presents an LLM Ensemble and Arbiter framework for cross-jurisdiction legal risk triage in digital media. The framework combines: (i) jurisdiction-specific legal grounding through a structured legal knowledge base (LawKB; 61 legal cards covering Ukrainian and EU norms), (ii) five heterogeneous open-weight LLMs, and (iii) a reproducible arbitration policy based on median risk aggregation and disagreement monitoring as a trigger for human review. The pipeline generates separate UA and EU risk assessments, a conservative global verdict, and a traceable rationales to support auditability.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>The framework was evaluated on 288 news items with lawyer validation across 576 jurisdiction-specific records. In the global mode, the system achieved a mean absolute error of 0.86 and a confidence-weighted mean absolute error of 0.63, with an approval rate of 95.8 and 95.1% accuracy in selecting the best-performing model rationale, while limiting missed CRITICAL-level cases to 6.3%.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Discussion<\/jats:title>\n                    <jats:p>These results indicate that the proposed workflow can support conservative, inspectable, and human-supervised legal risk triage across two jurisdictions. Rather than replacing legal judgment, the framework functions as a preliminary decision-support mechanism that strengthens transparency, structured escalation, auditability, and procedural accountability in high-risk digital-media environments.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/frai.2026.1841104","type":"journal-article","created":{"date-parts":[[2026,7,2]],"date-time":"2026-07-02T05:33:31Z","timestamp":1782970411000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":1,"title":["Bridging jurisdictions and legislatures: an LLM ensemble and arbiter framework for automated legal risk triage in digital media"],"prefix":"10.3389","volume":"9","author":[{"given":"Khrystyna","family":"Lipianina-Honcharenko","sequence":"first","affiliation":[{"name":"Department for Information Computer Systems and Control, West Ukrainian National University","place":["Ternopil, Ukraine"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tetiana","family":"Drakokhrust","sequence":"additional","affiliation":[{"name":"Department of Theory of Law and Constitutionalism, West Ukrainian National University","place":["Ternopil, Ukraine"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Pavlo","family":"Bykovyy","sequence":"additional","affiliation":[{"name":"Department for Information Computer Systems and Control, West Ukrainian National University","place":["Ternopil, Ukraine"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Kyrylo","family":"Turchynov","sequence":"additional","affiliation":[{"name":"Department of Theory of Law and Constitutionalism, West Ukrainian National University","place":["Ternopil, Ukraine"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ihor","family":"Ihnatiev","sequence":"additional","affiliation":[{"name":"Department of Cyber Security, West Ukrainian National University","place":["Ternopil, Ukraine"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,7,2]]},"reference":[{"key":"ref1","volume-title":"Artificial Intelligence Risk Management Framework: Generative Artificial Intelligence Profile (NIST AI 600\u20131)","author":"Autio","year":"2024"},{"key":"ref2","author":"Bohuta","year":"2025"},{"key":"ref3","doi-asserted-by":"crossref","DOI":"10.18653\/v1\/2020.findings-emnlp.261","article-title":"LEGAL-BERT: the Muppets straight out of law school","author":"Chalkidis","year":"2020"},{"key":"ref4","doi-asserted-by":"crossref","DOI":"10.2139\/ssrn.3936759","article-title":"LexGLUE: a benchmark dataset for legal language understanding in English","author":"Chalkidis","year":"2021"},{"key":"ref5","volume-title":"Convention for the Protection of Human Rights and Fundamental Freedoms","year":"1950"},{"key":"ref6","first-page":"55","volume-title":"Council Framework Decision 2008\/913\/JHA of 28 November 2008 on Combating Certain Forms and Expressions of Racism and Xenophobia by means of criminal law. Luxembourg: Publications Office of the European Union. OJ L 328, 2008","year":"2008"},{"key":"ref7","article-title":"DeepSeek-R1-distill-llama-8B: incentivizing reasoning capability in LLMs via reinforcement learning","year":"2024"},{"key":"ref8","doi-asserted-by":"publisher","first-page":"1977","DOI":"10.1057\/s41599-025-05924-3","article-title":"Large language models in legal systems: a survey","volume":"12","author":"Dehghani","year":"2025","journal-title":"Hum. Soc. Sci. Commun."},{"key":"ref9","volume-title":"Proposal for a Regulation of the European Parliament and of the Council Laying down Harmonised Rules on Artificial Intelligence (Artificial Intelligence Act) and Amending Certain Union Legislative Acts. Brussels: COM (2021) 206 final","year":"2021"},{"key":"ref9001","first-page":"22","article-title":"Directive 2005\/29\/EC of 11 May 2005 concerning unfair business-to-consumer commercial practices in the internal market","year":"2005"},{"key":"ref10","first-page":"1","volume-title":"Directive 2011\/93\/EU of 13 December 2011 on Combating the Sexual Abuse and Sexual Exploitation of Children and Child Pornography, and Replacing Council Framework Decision 2004\/68\/JHA. Luxembourg: Publications Office of the European Union, OJ L 335","year":"2011"},{"key":"ref11","first-page":"1","volume-title":"Regulation (EU) 2016\/679 of 27 April 2016 on the Protection of Natural Persons with Regard to the Processing of Personal Data and on the Free Movement of Such Data, and Repealing Directive 95\/46\/EC (General Data Protection Regulation)","year":"2016"},{"key":"ref12","first-page":"6","volume-title":"Directive (EU) 2017\/541 of 15 March 2017 on Combating Terrorism and Replacing Council Framework Decision 2002\/475\/JHA and Amending Council Decision 2005\/671\/JHA. OJ L 88","year":"2017"},{"key":"ref13","first-page":"1","volume-title":"Regulation (EU) 2022\/2065 of 19 October 2022 on a Single Market for Digital Services and Amending Directive 2000\/31\/EC (Luxembourg: Digital Services Act). OJ L 277","year":"2022"},{"key":"ref14","volume-title":"Regulation (EU) 2024\/1689 of 13 June 2024 Laying down Harmonised rules on artificial Intelligence and Amending Certain Union Legislative Acts (Luxembourg: Artificial Intelligence Act). OJ L 2024\/1689","year":"2024"},{"key":"ref15","volume-title":"Gemma (Including Gemma 2 2B-IT)","year":"2024"},{"key":"ref16","doi-asserted-by":"crossref","DOI":"10.52202\/075280-1915","article-title":"LegalBench: a collaboratively built benchmark for measuring legal reasoning in large language models","author":"Guha","year":"2023"},{"key":"ref17","volume-title":"Harvey [Software]","year":"2024"},{"key":"ref18","volume-title":"ISO 31022: Risk Management\u2014Guidelines for the Management of Legal Risk","year":"2020"},{"key":"ref19","article-title":"Leveraging passage retrieval with generative models for open-domain question answering","author":"Izacard","year":"2020"},{"key":"ref20","doi-asserted-by":"crossref","DOI":"10.1145\/3571730","article-title":"Survey of hallucination in natural language generation","author":"Ji","year":"2023"},{"key":"ref21","first-page":"611","author":"Kwon","year":"2023"},{"key":"ref22","article-title":"Retrieval-augmented generation for knowledge-intensive NLP tasks","author":"Lewis","year":"2020"},{"key":"ref23","volume-title":"Lexis+ AI","year":"2024"},{"key":"ref24","article-title":"Holistic evaluation of language models","author":"Liang","year":"2022"},{"key":"ref25","author":"Lipianina-Honcharenko","year":"2026"},{"key":"ref26","doi-asserted-by":"publisher","first-page":"202","DOI":"10.32620\/reks.2025.3.15","article-title":"A general method for real-time detection of information threats with a Ukraine case study","volume":"2025","author":"Lipianina-Honcharenko","year":"2025","journal-title":"Radioelectron. Comput. Syst."},{"key":"ref27","article-title":"SelfCheckGPT: zero-resource black-box hallucination detection for generative LLMs","author":"Manakul","year":"2023"},{"key":"ref28","volume-title":"Phi-3.5-mini-instruct","year":"2024"},{"key":"ref29","volume-title":"Mistral-7B-Instruct-v0.3","year":"2024"},{"key":"ref30","year":"2026"},{"key":"ref31","article-title":"LegalBench-RAG: benchmarking retrieval-augmented generation for legal tasks","author":"Pipitone","year":"2024"},{"key":"ref32","volume-title":"Qwen2.5-7B-Instruct","year":"2024"},{"key":"ref33","doi-asserted-by":"crossref","DOI":"10.6028\/NIST.AI.100-1","volume-title":"Artificial Intelligence Risk Management Framework (AI RMF 1.0)","author":"Tabassi","year":"2023"},{"key":"ref34","year":""},{"key":"ref35","year":""},{"key":"ref36","volume-title":"Vidomosti Verkhovnoi Rady Ukrainy (VVR), No. 48, Art. 650","year":"1992"},{"key":"ref37","volume-title":"Vidomosti Verkhovnoi Rady Ukrainy (VVR), No. 16, Art. 93","year":"1994"},{"key":"ref38","volume-title":"Vidomosti Verkhovnoi Rady Ukrainy (VVR), No. 30, Art. 141","year":"1996"},{"key":"ref39","volume-title":"Vidomosti Verkhovnoi Rady Ukrainy (VVR), No. 25\u201326, Art. 131","year":"2001"},{"key":"ref40","volume-title":"Vidomosti Verkhovnoi Rady Ukrainy (VVR), No. 34, Art. 481","year":"2010"},{"key":"ref41","volume-title":"Vidomosti Verkhovnoi Rady Ukrainy (VVR), 2023, No. 47\u201350, Art. 120","year":"2022"},{"key":"ref42","volume-title":"OpenAI-Compatible Server","year":"2026"},{"key":"ref43","article-title":"Self-consistency improves chain of thought reasoning in language models","author":"Wang","year":"2022"},{"key":"ref44","article-title":"AutoGen: enabling next-gen LLM applications via multi-agent conversation","author":"Wu","year":"2023"},{"key":"ref45","article-title":"ReAct: synergizing reasoning and acting in language models","author":"Yao","year":"2022"}],"container-title":["Frontiers in Artificial Intelligence"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2026.1841104\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,2]],"date-time":"2026-07-02T05:33:35Z","timestamp":1782970415000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2026.1841104\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,7,2]]},"references-count":46,"alternative-id":["10.3389\/frai.2026.1841104"],"URL":"https:\/\/doi.org\/10.3389\/frai.2026.1841104","relation":{},"ISSN":["2624-8212"],"issn-type":[{"value":"2624-8212","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,7,2]]},"article-number":"1841104"}}