How LexisNexis CTO Greg Dickason Tempers AI For Real-World Results
InformationWeek, Friday, July 17th, 2026
LexisNexis pairs human attorney review with competitive model testing to cut AI hallucinations in legal research.
LexisNexis is applying rigorous validation to ensure AI reliability in the legal sector, where accuracy is critical.
CTO Greg Dickason describes using models to test each other and enlisting a couple hundred lawyers to assess AI output before deployment.
Rather than relying on generic large language models, the company distills thousands of case files into concentrated research material so the model is not overwhelmed by excess information.
About 30% of the LexisNexis ecosystem is currently allocated to AI development, a share expected to grow substantially.
Dickason warns against using general-purpose AI tools for specialized legal work, citing the serious consequences of fabricated case citations appearing in court briefs.