A recent survey has revealed that while a significant majority of federal judges have employed generative artificial intelligence (AI) tools in their work, its daily application remains uncommon. This unprecedented survey, conducted by Northwestern University researchers in collaboration with the New York City Bar Association, provides new insights into the prevalence and patterns of AI usage within the federal judiciary.
Despite over 60% of judges having tried AI tools at least occasionally, only 5.4% reported using them daily. A further 17% use AI weekly, while a notable 38.4% have never used such technologies in their judicial duties. This disparity suggests that although AI is making inroads into federal court chambers, it has not yet become a standard component of decision-making processes among judges, as detailed in the research report.
Judges showed a marked preference for legal-specific AI tools over general-purpose systems. Westlaw AI-Assisted Research led the pack with a 38.4% usage rate among judges, followed by ChatGPT at 28.6%. Moreover, AI is predominantly utilized in legal research, with 30% employing it for this purpose, while applications like document review and drafting non-case documents remain secondary.
The study also shed light on AI usage by judicial staff, illustrating that 50.9% of judges do not use AI, whereas slightly fewer, 45%, reported that their staff refrained from utilizing AI as well. For staff, legal research takes precedence at 39.8%, mirroring the priorities and concerns of the judges themselves.
Reflecting concerns over the readiness of the judiciary for AI, nearly half of the judges reported a lack of AI training opportunities, despite an evident interest in receiving such training. Moreover, chambers exhibit varied governance strategies for AI use, with no single approach dominating, pointing toward an emerging area of policy development.
Beyond professional use, approximately 38% of judges engage with AI in their personal activities weekly or daily, citing a range of domestic applications from trip planning to personal correspondence drafting. This personal engagement could potentially influence professional adoption rates over time.
The survey revealed that federal judges have mixed feelings about AI’s future within the judiciary, with opinions split nearly evenly between optimism and concern. Respondents highlighted potential benefits like efficiency and better resource management, counterbalanced by worries over accuracy and the risk of reducing critical skill levels. This echoes broader societal debates concerning AI’s resilience and reliability.
This survey is the first of its kind to empirically evaluate AI’s penetration in federal courts, involving a random sample of various types of judges, including bankruptcy and district court judges. Nevertheless, due to the modest response rate and certain categories being underrepresented, these findings should be interpreted as indicative rather than conclusive.
Although this study represents a step toward understanding the judiciary’s evolving interaction with AI, future research will be crucial in navigating the complexities and challenges of integrating AI within judicial frameworks. For a more detailed analysis of these findings, please explore the full dataset and methodologies as published by the research team.