{"693637":{"#nid":"693637","#data":{"type":"news","title":"Researchers Create Tool to Detect AI-Hallucinated Legal Citations ","body":[{"value":"\u003Cp\u003EAs more legal documents are drafted with large language models (LLMs), falsified citations are becoming a problem for the people who rely on them. Georgia Tech researchers have created a tool to help lawyers verify their work.\u0026nbsp;\u0026nbsp;\u003C\/p\u003E\u003Cp\u003EThrough a donation from IBM, Georgia Tech\u2019s Center for Scientific Software Engineering (CSSE) and Open Source Program Office (OSPO) held a fellowship competition for students to work with mentors at CSSE and OSPO to create the tool. Master\u2019s students \u003Ca href=\u0022https:\/\/adrian123l.github.io\/\u0022\u003E\u003Cstrong\u003EAdrian Lozada\u003C\/strong\u003E\u003C\/a\u003E and \u003Ca href=\u0022https:\/\/woodygoodenough.com\/\u0022\u003E\u003Cstrong\u003EDi Wu\u003C\/strong\u003E\u003C\/a\u003E were selected as fellows to work on the project.\u0026nbsp;\u0026nbsp;\u003C\/p\u003E\u003Cp\u003E\u003Ca href=\u0022https:\/\/github.com\/gt-csse\/mellea-lrc\/\u0022\u003EThe Mellea Legal Requirements Checker\u003C\/a\u003E (Mellea-LRC) can quickly identify and accurately verify case law citations, specifically checking for citations that are hallucinated by artificial intelligence (AI). One fellowship goal was to give lawyers and others in the legal field a tool to easily verify legal citations. Mella-LRC does this by using both LLMs and IBM\u2019s Mellea open-source framework to verify and repair LLM prompts. \u0026nbsp;\u003C\/p\u003E\u003Cp\u003EThe team tested the tool on 26 public court-filing documents containing falsified citations. Mellea-LRC can process a document in about eight minutes. The tool first extracts the case law citations from the legal document. It then verifies that each citation is authentic and supports the document\u0027s content.\u202f\u0026nbsp;\u003C\/p\u003E\u003Cp\u003E\u201cIBM\u2019s Mellea framework was the foundation we built it on. Mellea\u2019s job is to make model calls predictable: you write the rules the output must satisfy as ordinary code, and the framework retries until LLM output meets them. Our job was to leverage the framework to build the legal application from end to end,\u201d Wu said.\u0026nbsp;\u0026nbsp;\u003C\/p\u003E\u003Cp\u003EAn important aspect of the project was ensuring the tool\u2019s code was open source. The team supported this goal by also creating and releasing a new test and validation dataset on Hugging Face \u2013 a popular open-source AI development community \u2013 for future benchmarking studies.\u0026nbsp;\u003C\/p\u003E\u003Cp\u003E\u201cAgentic AI is growing quickly, and much of the growth happens in open-source projects, where anyone can read, use, and improve the work,\u201d Lozada said.\u0026nbsp;\u0026nbsp;\u003C\/p\u003E\u003Cp\u003EOSPO Director and Principal Research Scientist \u003Ca href=\u0022https:\/\/jyoung3131.github.io\/\u0022\u003E\u003Cstrong\u003EJeff\u003C\/strong\u003E \u003Cstrong\u003EYoung\u003C\/strong\u003E\u003C\/a\u003E noted that IBM\u2019s Mellea framework was critical to this work in that its \u201cinstruct, verify, repair\u201d paradigm allows for the creation of reliable LLM loops as well as the use of small language models (SLMs), such as IBM\u2019s Granite, which can be run on a laptop.\u0026nbsp;\u003C\/p\u003E\u003Cp\u003E\u201cThis project demonstrates the power of open-source models like IBM Granite as well as the importance of using novel software engineering and algorithmic techniques from the Mellea project to improve the reliability of our existing LLMs and SLMs.\u0026nbsp;\u0026nbsp;\u003C\/p\u003E\u003Cp\u003E\u201cMellea-LRC provides a compelling example of AI\u2019s potential to evolve from a source of error in the legal profession into a trusted ally by helping lawyers work faster and cite precedents more accurately,\u201d Young said.\u0026nbsp;\u0026nbsp;\u003C\/p\u003E","summary":"","format":"limited_html"}],"field_subtitle":"","field_summary":[{"value":"\u003Cdiv\u003E\u003Cp\u003EThrough a donation from IBM, Georgia Tech\u2019s Center for Scientific Software Engineering (CSSE) and Open Source Program Office (OSPO) held a fellowship competition for students to work with mentors at CSSE and OSPO to create the tool. Master\u2019s students \u003Cstrong\u003EAdrian Lozada\u003C\/strong\u003E and \u003Cstrong\u003EDi Wu\u003C\/strong\u003E were selected as fellows to work on the project.\u0026nbsp;\u0026nbsp;\u003C\/p\u003E\u003C\/div\u003E\u003Cdiv\u003E\u003Cp\u003EThe Mellea Legal Requirements Checker (Mellea-LRC) can quickly identify and accurately verify case law citations, specifically checking for citations that are hallucinated by artificial intelligence (AI). One fellowship goal was to give lawyers and others in the legal field a tool to easily verify legal citations. Mella-LRC does this by using both LLMs and IBM\u2019s Mellea open-source framework to verify and repair LLM prompts. \u0026nbsp;\u003C\/p\u003E\u003C\/div\u003E","format":"limited_html"}],"field_summary_sentence":[{"value":"Georgia Tech researchers have created a tool that detects AI-hallucinated citations in legal documents. "}],"uid":"36532","created_gmt":"2026-10-29 21:10:19","changed_gmt":"2026-10-30 19:37:40","author":"Morgan Usry","boilerplate_text":"","field_publication":"","field_article_url":"","location":"Atlanta, GA","dateline":{"date":"2026-10-30T00:00:00-04:00","iso_date":"2026-10-30T00:00:00-04:00","tz":"America\/New_York"},"extras":[],"hg_media":{"681600":{"id":"681600","type":"image","title":"AdobeStock_2cc0dae5-a93c-42b2-bf30-419166aa8f0b.jpg","body":null,"created":"1793308229","gmt_created":"2026-10-29 21:10:29","changed":"1793308229","gmt_changed":"2026-10-29 21:10:29","alt":"AI analyzing legal document","file":{"fid":"266026","name":"AdobeStock_2cc0dae5-a93c-42b2-bf30-419166aa8f0b.jpg","image_path":"\/sites\/default\/files\/2026\/10\/29\/AdobeStock_2cc0dae5-a93c-42b2-bf30-419166aa8f0b.jpg","image_full_path":"http:\/\/hg.gatech.edu\/\/sites\/default\/files\/2026\/10\/29\/AdobeStock_2cc0dae5-a93c-42b2-bf30-419166aa8f0b.jpg","mime":"image\/jpeg","size":107712,"path_740":"http:\/\/hg.gatech.edu\/sites\/default\/files\/styles\/740xx_scale\/public\/2026\/10\/29\/AdobeStock_2cc0dae5-a93c-42b2-bf30-419166aa8f0b.jpg?itok=vOEwx20X"}}},"media_ids":["681600"],"groups":[{"id":"47223","name":"College of Computing"},{"id":"50875","name":"School of Computer Science"}],"categories":[{"id":"194606","name":"Artificial Intelligence"},{"id":"153","name":"Computer Science\/Information Technology and Security"},{"id":"135","name":"Research"},{"id":"8862","name":"Student Research"}],"keywords":[{"id":"2556","name":"artificial intelligence"},{"id":"5155","name":"open source"}],"core_research_areas":[{"id":"193655","name":"Artificial Intelligence at Georgia Tech"}],"news_room_topics":[],"event_categories":[],"invited_audience":[],"affiliations":[],"classification":[],"areas_of_expertise":[],"news_and_recent_appearances":[],"phone":[],"contact":[{"value":"\u003Cp\u003EMorgan Usry, Communications Officer, School of Computer Science\u003C\/p\u003E","format":"limited_html"}],"email":["morgan.usry@cc.gatech.edu"],"slides":[],"orientation":[],"userdata":""}}}