by Dennis Crouch
I enjoyed reading Professor S.I. Strong’s new article on AI-lawyering and her proposed solution that inspired by the English legal profession’s structure. Although I think this is all still early stage, emerging empirical research is showing how generative AI usage often triggers “metacognitive laziness” and “cognitive offloading” in users. Studies particularly with law students show reduced motivation to learn, diminished ability to self-regulate, and less deep engagement with material—gaining only improved short-term performance on individual tasks. Routine reliance on AI usage is also showing an atrophy of other critical skills. Namely, when professionals have confidence in GenAI’s ability to perform a task, those folks spend much less effort thinking critically about the issues.
One solution in the article is derived from Strong’s experience as both a U.S. attorney and a U.K. solicitor. She proposes dividing the legal profession into two groups: “post-AI solicitors,” who would provide legal services with robust AI assistance handling standardized, routine matters; and “post-AI barristers,” who would work without reliance on generative AI, focusing on novel or complex questions of law. Post-AI solicitors would undertake more vocational training focused on proper AI use, while post-AI barristers would focus on traditional legal research and drafting without AI assistance. Strong argues this approach would maintain public confidence in the legal profession and justice system, as humans would still independently analyze difficult legal questions while still permitting the efficiency gains and potential access-to-justice that comes with the automated tools.
While thought provoking, I don’t believe we’re ready for fundamental structural change to the level Strong proposes here. AI continues to evolve rapidly, and we’re nowhere near a settled state regarding what tools lawyers are or will be using even within the typical three-year law school timeframe. I acknowledge that concerns about “metacognitive laziness” and “cognitive offloading” are genuine—but these issues aren’t unique to AI. Similar dynamics regularly emerge when experienced attorneys begin to rely heavily upon paralegals or secretaries who have proven their worth. On this point though, AI does run the risk of pushing this phenomenon to much higher level tasks than we’ve seen before, potentially affecting core analytical and reasoning skills that have traditionally defined the legal profession.
Many of us are also working on more direct ways to address the cognitive challenges posted by GenAI without entirely rejecting technological tools. Currently, mindful AI usage largely depends on personal discipline—similar to how some individuals successfully limit their phone scrolling habits. However, we’re working toward both technological and organizational mechanisms that can help structure AI interactions to promote cognitive enhancement rather than detriment. These approaches often focus on “cognitive forcing,” although my approach would be “curiosity-promoting” interfaces that prompt deeper engagement rather than passive consumption.