The legal industry has broadly embraced artificial intelligence, but the pace of adoption is running well ahead of the professional judgment needed to use it responsibly.
Comfort with AI tools has grown rapidly, with lawyers and legal professionals now routinely turning to them for tasks ranging from polishing emails to summarising documents and generating strategic outlines.
Yet that comfort, according to legal innovator Brendan W. Miller, J.D., carries a significant catch that the profession cannot afford to ignore.
The landmark case of *Mata v. Avianca, Inc.* remains the most cited cautionary tale, in which attorneys were sanctioned for submitting AI-generated hallucinated case citations to the court.
Miller argues the core lesson from that case is widely misunderstood, noting that “the core issue wasn’t that AI was used, but rather that AI was not properly supervised.”
The duty of verification, courts have made clear, is one that cannot be delegated to any technology, regardless of how sophisticated or widely adopted it becomes.
Professional liability insurance frameworks are also shifting in response, with emerging guidance warning lawyers explicitly against relying on unverified AI outputs in their work.
Miller describes certain tasks and decisions as reflecting “the exercise of professional responsibility and the very essence of client service,” adding they “demand the autonomous exercise of professional judgment” and “cannot and should not be delegated to AI.”
The legal industry has moved through an initial phase focused on gaining access to AI tools and standing up practical use cases across firms and legal departments.
Miller contends the profession has now entered a second and more consequential phase, one centred on judgment, defensibility, and the responsible governance of AI-assisted work.
Courts have already sanctioned AI misuse, bar associations have clarified overriding professional duties, and clients are increasingly asking pointed questions about how their legal service providers govern AI use internally.
The organisations best positioned to lead in this environment, Miller argues, are those that move beyond training legal professionals to prompt AI better, and instead focus on teaching them how to decide better.
That distinction, between mechanical skill with AI tools and mature judgment about when and how to rely on them, is becoming a defining competitive and ethical line within the legal industry.
Miller, a litigator, corporate attorney, and legal technologist, frames legal innovation as “continually being relevant for clients, by making the business and practice of law easier, better, and more valuable.”
Building AI literacy through scenario-based training is among the emerging strategies legal organisations are beginning to adopt as they work to close the gap between adoption speed and judgment maturity.
The broader message for the legal profession is clear: the question is no longer whether to use AI, but whether the humans overseeing it are genuinely equipped to exercise the judgment that professional responsibility demands.