The three-bucket framework to navigating AI in the courts
Watch now
© 2025 Thomson Reuters TR6363228/11-25
Artificial Intelligence (AI) is reshaping how trial courts function, streamlining workflows, improving efficiency, and introducing new challenges. As legal professionals increasingly adopt AI tools, it’s critical to understand both their capabilities and limitations. Thoughtful integration and vigilant oversight are essential to uphold justice and maintain the integrity of the judicial process.
This topic was recently explored in the ‘Clarity on Tech’ podcast series, where it is examined how AI is currently being used in trial courts and what its future might hold.
Building on that conversation, we’ve outlined below a three-section framework to strengthen justice without compromising integrity:
of survey respondents said their court was currently using GenAI
17%
"AI won't replace humans, but humans using AI will replace those that aren't."
Anticipated individual time savings up to
hours per week within five years.
8.8
Hear from industry experts how courts can navigate both transformative opportunities and challenges of AI
Watch the on-demand webinar to hear benefits, risks, and strategies of professional-grade AI solutions to meet today's challenges and prepare for tomorrow's opportunities.
(Source: Transforming state courts: Harnessing AI for operational efficiency and access to justice)
- Tina Tierney, Esq., MBA, Senior Client Executive, Thomson Reuters
Explore how courts can responsibly implement AI to enhance efficiency, uphold justice, and maintain public confidence.
(Source: Staffing, Operations and Technology: A 2025 survey of State Courts)
(Source: Transforming state courts: Harnessing AI for operational efficiency and access to justice)
Lorem Ipsum
Lorem Ipsum
Bucket 3: Use
Bucket 2: Outputs
Bucket 1: Inputs
Protect sensitive information
• Ensure the technology you’re using has strict security measures
• Implement data governance protocols for AI systems
• Train staff on what information can and cannot be shared with AI tools
Quality data management
• Verify accuracy of information before inputting
• Review all information for potential bias
• Create standardized input procedures across court functions
What are we putting into the system?
Verification is critical
• Learn what AI "hallucinations" are and techniques to spot and verify them • Always verify AI-generated citations and legal references • Cross-check outputs against authoritative sources • Quality assessment • Evaluate AI outputs for bias, accuracy, and relevance concerns • Establish benchmarks for acceptable AI performance • Create feedback loops to improve AI system reliability
What's coming out of the system?
What are AI hallucinations?
Hallucinations occur when AI generates false information and presents it as fact.
Never replace human judgment
• AI should never make court decisions independently • Maintain thorough human review for all outputs • Use AI as a "thought partner," not a decision-maker
Suitable applications
• Administrative tasks: Case management, motion prioritization, scheduling optimization • Research support: Understanding complex technical information, legal research assistance • Self-represented litigant aid: Providing process guidance and form assistance • Preparation tools: Brief drafting support, hearing preparation, presentation creation
How are we going to use this information?
Bucket 3: Use
"Artificial intelligence should never be making a decision in a court. It can be like you said a thought tool in research, but it's not a replacement for our research and our analysis."
- Judge Pamela Gates
Bucket 2: Outputs
"It gave me five citations. The first one I knew because I had found that and referenced it and the other four were hallucinations."
- Judge Marissa Brasswell
Bucket 1: Inputs
"Anyone who will be using artificial intelligence within our court will need to take training on all of the limitations and risks associated with artificial intelligence."
- Judge Pamela Gates
Lorem Ipsum
Lorem Ipsum
Lorem Ipsum
How are we going to use this information?
Never replace human judgment
• AI should never make court decisions independently • Maintain thorough human review for all outputs • Use AI as a "thought partner," not a decision-maker
Suitable applications
• Administrative tasks: Case management, motion
prioritization, scheduling optimization • Research support: Understanding complex technical information, legal research assistance • Self-represented litigant aid: Providing process
guidance and form assistance • Preparation tools: Brief drafting support, hearing
preparation, presentation creation
"Artificial intelligence should never be making a decision in a court. It can be like you said a thought tool in research, but it's not a replacement for our research and our analysis."
- Judge Pamela Gates
How are we going to use this information?
Never replace human judgment
• AI should never make court decisions independently • Maintain thorough human review for all outputs • Use AI as a "thought partner," not a decision-maker
Suitable applications
• Administrative tasks: Case management, motion
prioritization, scheduling optimization • Research support: Understanding complex technical information, legal research assistance • Self-represented litigant aid: Providing process
guidance and form assistance • Preparation tools: Brief drafting support, hearing
preparation, presentation creation
"Artificial intelligence should never be making a decision in a court. It can be like you said a thought tool in research, but it's not a replacement for our research and our analysis."
- Judge Pamela Gates