Thinking Like A Lawyer In The Age Of Generative AI: Cognitive Limits On AI Adoption Among Lawyers

Abstract

As of mid-2025, there is robust evidence that generative AI possesses the technological capability to significantly reshape legal practice. Yet legal markets and doctrines have, to date, remained largely unchanged. This gap is often attributed to familiar lags in technology adoption and a range of socio-legal and institutional constraints. This Essay offers a complementary explanation: lawyers face distinctive and under-appreciated challenges to deeply understanding AI-generated outputs to complex or unfamiliar legal problems. These difficulties, which have nothing to do with "hallucinated" sources, hamper lawyers' ability to evaluate the quality of AI assistance and to perform closely related tasks-such as engaging with clients or judges, tailoring arguments to new contexts, or synthesizing insights across legal issues. Although these limitations may diminish as the profession adapts, they may also reflect more fundamental features of human legal reasoning, particularly among junior and less experienced attorneys. If so, these dynamics are likely to influence AI's role in legal practice for the foreseeable future, while also offering critical insights into how lawyers ought to engage with AI tools.