{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,23]],"date-time":"2026-09-23T03:37:35Z","timestamp":1790134655641,"version":"4.0.1"},"reference-count":39,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,3,26]],"date-time":"2026-03-26T00:00:00Z","timestamp":1774483200000},"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:p>\n                    The growing demand for automation in legal contract management exposes a persistent limitation of current language models: insufficient adaptation to the semantic, structural, and regulatory constraints of legal language. While large language models perform well on general NLP tasks, their direct application to legal document classification, clause extraction, and contract summarization often yields unstable, legally unreliable outputs. This work presents a structured methodological pipeline for evaluating and adapting language models for legal contract automation, combining domain-specific fine-tuning of open-source models with a controlled comparative assessment against large general-purpose LLMs used exclusively in inference mode. The methodology integrates legal corpus curation, clause-level annotation, and efficient adaptation techniques, and is evaluated across three core tasks: contract document classification, normative clause extraction, and regulatory summarization. The evaluation protocol is explicitly designed to disentangle the effects of supervision from deployment constraints arising in regulated legal settings. Experimental results show consistent and statistically significant performance gains for legally adapted models over general-purpose baselines, achieving Macro-F1 of 0.921 in classification, span-level F1 of 0.903 in clause extraction, and ROUGE-L of 0.886 in summarization (\n                    <jats:italic>p<\/jats:italic>\n                    &amp;lt; 0.01). Robustness analysis and cross-validation confirm stability across heterogeneous private-sector contract types. The findings should be interpreted under the evaluated comparison regime and highlight that, in legally constrained multi-stage workflows, task-aligned supervision provides measurable structural benefits that are not reducible to model scale alone when general-purpose LLMs are restricted to inference-only deployment.\n                  <\/jats:p>","DOI":"10.3389\/frai.2026.1782405","type":"journal-article","created":{"date-parts":[[2026,3,26]],"date-time":"2026-03-26T06:35:51Z","timestamp":1774506951000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":1,"title":["Technical evaluation of language models adapted for the automation of legal contracts: clause extraction, classification, and summarization"],"prefix":"10.3389","volume":"9","author":[{"given":"Jaime","family":"Govea","sequence":"first","affiliation":[{"name":"Escuela de Ingenier\u00eda en Ciberseguridad, Facultad de Ingenier\u00eda y Ciencias Aplicadas, Universidad de Las Am\u00e9ricas","place":["Quito, Ecuador"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Iv\u00e1n","family":"Ortiz-G\u00e1rces","sequence":"additional","affiliation":[{"name":"Escuela de Ingenier\u00eda en Ciberseguridad, Facultad de Ingenier\u00eda y Ciencias Aplicadas, Universidad de Las Am\u00e9ricas","place":["Quito, Ecuador"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Pablo","family":"Palacios","sequence":"additional","affiliation":[{"name":"Escuela de Inform\u00e1tica y Telecomunicaciones, Universidad Diego Portales","place":["Santiago, Chile"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Alexandra Maldonado","family":"Navarro","sequence":"additional","affiliation":[{"name":"Escuela de Posgrados, Maestr\u00eda en Derecho Digital, Universidad de Las Am\u00e9ricas","place":["Quito, Ecuador"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Santiago Acurio","family":"Del Pino","sequence":"additional","affiliation":[{"name":"Facultad de Derecho y Sociedad, Pontificia Universidad Cat\u00f3lica del Ecuador","place":["Quito, Ecuador"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"William","family":"Villegas-Ch","sequence":"additional","affiliation":[{"name":"Escuela de Ingenier\u00eda en Ciberseguridad, Facultad de Ingenier\u00eda y Ciencias Aplicadas, Universidad de Las Am\u00e9ricas","place":["Quito, Ecuador"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,3,26]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2022.emnlp-main.519","article-title":"\u201cEUR-Lex-Sum: a multi- and cross-lingual dataset for long-form summarization in the legal domain,\u201d","author":"Aumiller","year":"2022","journal-title":"Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing"},{"key":"B2","doi-asserted-by":"publisher","DOI":"10.1145\/3442188.3445922","article-title":"\u201cOn the dangers of stochastic parrots,\u201d","author":"Bender","year":"2021","journal-title":"Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency"},{"key":"B3","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1145\/3628602","article-title":"Measuring and mitigating gender bias in legal contextualized language models","volume":"18","author":"Bozdag","year":"2024","journal-title":"ACM Trans. Knowl. Discov. Data"},{"key":"B4","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2020.findings-emnlp.261","article-title":"\u201cLEGAL-BERT: the muppets straight out of law school,\u201d","author":"Chalkidis","year":"2020","journal-title":"Findings of the Association for Computational Linguistics: EMNLP 2020"},{"key":"B5","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2022.acl-long.297","article-title":"\u201cLexGLUE: a benchmark dataset for legal language understanding in English,\u201d","author":"Chalkidis","year":"2022","journal-title":"Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)"},{"key":"B6","doi-asserted-by":"publisher","DOI":"10.1145\/3485447.3511998","article-title":"\u201cKnowPrompt: knowledge-aware prompt-tuning with synergistic optimization for relation extraction,\u201d","author":"Chen","year":"2022","journal-title":"Proceedings of the ACM Web Conference 2022"},{"key":"B7","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2023.findings-acl.732","article-title":"\u201cFiDO: fusion-in-decoder optimized for stronger performance and faster inference,\u201d","author":"de Jong","year":"2023","journal-title":"Findings of the Association for Computational Linguistics: ACL 2023"},{"key":"B8","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/N19-1423","article-title":"\u201cBERT: pre-training of deep bidirectional transformers for language understanding,\u201d","author":"Devlin","year":"2019","journal-title":"Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT)"},{"key":"B9","article-title":"\u201cThe potential for jurisdictional challenges to AI or LLM training datasets,\u201d","author":"Draper","year":"2023","journal-title":"CEUR Workshop Proceedings"},{"key":"B10","doi-asserted-by":"publisher","first-page":"22167","DOI":"10.1109\/ACCESS.2024.3364533","article-title":"Leveraging NLP techniques for privacy requirements engineering in user stories","volume":"12","author":"Herwanto","year":"2024","journal-title":"IEEE Access"},{"key":"B11","article-title":"\u201cLoRA: low-rank adaptation of large language models,\u201d","author":"Hu","year":"2022","journal-title":"Proceedings of the 10th International Conference on Learning Representations (ICLR)"},{"key":"B12","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2023.nllp-1.22","article-title":"\u201cLarge language models are legal but they are not: making the case for a powerful LegalLLM,\u201d","author":"Jayakumar","year":"2023","journal-title":"Proceedings of the Natural Legal Language Processing Workshop 2023"},{"key":"B13","doi-asserted-by":"publisher","first-page":"100210","DOI":"10.1016\/j.caeai.2024.100210","article-title":"Fine-tuning ChatGPT for automatic scoring","volume":"6","author":"Latif","year":"2024","journal-title":"Comput. Educ"},{"key":"B14","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-72344-5_20","article-title":"\u201cCSAFT: continuous semantic augmentation fine-tuning for legal large language models,\u201d","author":"Li","year":"2024","journal-title":"Artificial Neural Networks and Machine Learning"},{"key":"B15","doi-asserted-by":"publisher","first-page":"621","DOI":"10.1007\/s11571-020-09646-x","article-title":"A novel index of functional connectivity: phase lag based on Wilcoxon signed rank test","volume":"15","author":"Li","year":"2020","journal-title":"Cogn. Neurodyn"},{"key":"B16","doi-asserted-by":"publisher","first-page":"105908","DOI":"10.1016\/j.clsr.2023.105908","article-title":"ITALIAN-LEGAL-BERT models for improving natural language processing tasks in the Italian legal domain","volume":"52","author":"Licari","year":"2024","journal-title":"Comput. Law Secur. Rev"},{"key":"B17","article-title":"\u201cCategory-extensible out-of-distribution detection via hierarchical context descriptions,\u201d","author":"Liu","year":"2023","journal-title":"Advances in Neural Information Processing Systems (NeurIPS 2023)"},{"key":"B18","doi-asserted-by":"publisher","first-page":"012022","DOI":"10.1088\/1742-6596\/1840\/1\/012022","article-title":"Information technology for mobile perimeter security systems creation","volume":"1840","author":"Lobanchykova","year":"2021","journal-title":"J. Phys"},{"key":"B19","doi-asserted-by":"publisher","DOI":"10.1109\/ICDMW60847.2023.00037","article-title":"\u201cRacial skew in fine-tuned legal AI language models,\u201d","author":"Malic","year":"2023","journal-title":"2023 IEEE International Conference on Data Mining Workshops (ICDMW)"},{"key":"B20","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.nllp-1.11","article-title":"\u201cEnhancing contract negotiations with LLM-based legal document comparison,\u201d","author":"Narendra","year":"2024","journal-title":"Proceedings of the Natural Legal Language Processing Workshop 2024"},{"key":"B21","doi-asserted-by":"publisher","first-page":"99","DOI":"10.14718\/NovumJus.2023.17.1.4","article-title":"Policies and guidelines for the implementation of clear and easy legal language","volume":"17","author":"Ortega-Ruiz","year":"2023","journal-title":"Novum Jus"},{"key":"B22","doi-asserted-by":"publisher","first-page":"76","DOI":"10.1017\/err.2024.99","article-title":"Potentials and challenges of large language models (LLMs) in the context of administrative decision-making","volume":"16","author":"Pesch","year":"2025","journal-title":"Eur. J. Risk Regul"},{"key":"B23","doi-asserted-by":"crossref","first-page":"493","DOI":"10.1007\/978-981-19-0840-8_37","article-title":"\u201cLegal text analysis using pre-trained transformers,\u201d","volume-title":"Advanced Machine Intelligence and Signal Processing","author":"Prajwal","year":"2022"},{"key":"B24","article-title":"\u201cFintech and Tokenization: a legislative study in Argentina and Spain about the application of Blockchain in the field of properties,\u201d","author":"Schmidt","year":"2020","journal-title":"ADCAIJ: Advances in Distributed Computing and Artificial Intelligence Journal"},{"key":"B25","doi-asserted-by":"publisher","first-page":"519","DOI":"10.1007\/s10506-024-09394-x","article-title":"Legal sentence boundary detection using hybrid deep learning and statistical models","volume":"33","author":"Sheik","year":"2024","journal-title":"Artif. Intell. Law"},{"key":"B26","doi-asserted-by":"publisher","first-page":"18253","DOI":"10.1109\/ACCESS.2025.3533217","article-title":"Exploring LLMs applications in law: a literature review on current legal NLP approaches","volume":"13","author":"Siino","year":"2025","journal-title":"IEEE Access"},{"key":"B27","doi-asserted-by":"publisher","first-page":"8954","DOI":"10.1080\/03610926.2024.2315297","article-title":"On integer partitions and the Wilcoxon rank-sum statistic","volume":"53","author":"Sills","year":"2024","journal-title":"Commun. Stat. Theory Methods"},{"key":"B28","first-page":"130","article-title":"\u201cExploring domain and task adaptation of LamBERTa models for article retrieval on the Italian civil code,\u201d","author":"Simeri","year":"2023","journal-title":"Proceedings of the Italian Research Conference on Digital Libraries (IRCDL)"},{"key":"B29","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2021.emnlp-main.160","article-title":"\u201cFast WordPiece tokenization,\u201d","author":"Song","year":"2021","journal-title":"Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing"},{"key":"B30","doi-asserted-by":"publisher","first-page":"417","DOI":"10.1007\/s10506-021-09301-8","article-title":"Unsupervised law article mining based on deep pre-trained language representation models with application to the Italian civil code","volume":"30","author":"Tagarelli","year":"2021","journal-title":"Artif. Intell. Law"},{"key":"B31","doi-asserted-by":"publisher","DOI":"10.3233\/FAIA240936","article-title":"\u201cFsPONER: few-shot prompt optimization for named entity recognition in domain-specific scenarios,\u201d","author":"Tang","year":"2024","journal-title":"ECAI 2024"},{"key":"B32","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2023.acl-short.30","article-title":"\u201cLanguage models get a gender makeover: mitigating gender bias with few-shot data interventions,\u201d","author":"Thakur","year":"2023","journal-title":"Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics"},{"key":"B33","doi-asserted-by":"publisher","DOI":"10.3233\/FAIA241266","article-title":"\u201cInSaAF: incorporating safety through accuracy and fairness - are LLMs ready for the Indian legal domain?,\u201d","author":"Tripathi","year":"2024","journal-title":"Proceedings of the International Conference on Legal Knowledge and Information Systems"},{"key":"B34","first-page":"1235","article-title":"\u201cLEDGAR: a large-scale multi-label corpus for text classification of legal provisions in contracts,\u201d","author":"Tuggener","year":"2020","journal-title":"Proceedings of the Twelfth Language Resources and Evaluation Conference"},{"key":"B35","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2023.findings-emnlp.831","article-title":"\u201cLLMs \u2013 the Good, the Bad or the Indispensable? A use case on legal statute prediction and legal judgment prediction on indian court cases,\u201d","author":"Vats","year":"2023","journal-title":"Findings of the Association for Computational Linguistics: EMNLP 2023"},{"key":"B36","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-36805-9_24","article-title":"\u201cJurisBERT: a new approach that converts a classification corpus into an STS one,\u201d","author":"Viegas","year":"2023","journal-title":"Computational Science and Its Applications"},{"key":"B37","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.findings-acl.333","article-title":"\u201cUnveiling selection biases: exploring order and token sensitivity in large language models,\u201d","author":"Wei","year":"2024","journal-title":"Findings of the Association for Computational Linguistics ACL 2024"},{"key":"B38","doi-asserted-by":"publisher","first-page":"100211","DOI":"10.1016\/j.hcc.2024.100211","article-title":"A survey on large language model (LLM) security and privacy: the good, the bad, and the ugly","volume":"4","author":"Yao","year":"2024","journal-title":"High-Confid. Comput"},{"key":"B39","doi-asserted-by":"publisher","DOI":"10.1145\/3462757.3466088","article-title":"\u201cWhen does pretraining help?,\u201d","author":"Zheng","year":"2021","journal-title":"Proceedings of the Eighteenth International Conference on Artificial Intelligence and Law"}],"container-title":["Frontiers in Artificial Intelligence"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2026.1782405\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,3,26]],"date-time":"2026-03-26T06:35:53Z","timestamp":1774506953000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2026.1782405\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,3,26]]},"references-count":39,"alternative-id":["10.3389\/frai.2026.1782405"],"URL":"https:\/\/doi.org\/10.3389\/frai.2026.1782405","relation":{},"ISSN":["2624-8212"],"issn-type":[{"value":"2624-8212","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,3,26]]},"article-number":"1782405"}}