Nearly four years into the generative AI (GenAI) boom, the legal industry is still figuring out how to teach legal judgement — the one skill no algorithm can replace. In fact, the state of AI in the legal industry and the implications for legal talent were not discussed in mainstream talent circles until 12 to 18 months ago, according to Lucie Allen, Co-CEO of legal talent and learning company Barbri.

Like many organizations, legal employers focused their early attention on tool adoption and workflow changes, Allen explains, but the harder question of how to build judgment in an AI-saturated environment has only recently moved to center stage.

Why AI cannot be bolted onto old legal models

Over the last five years, Allen says she's watched law firms and law schools scramble to adapt, and the biggest mistake they make is treating AI as an add-on to existing training. Instead, they should be completely rethinking lawyer training and education from the ground up. To do so, lawyer competencies must become the foundation, and then legal organizations need to look "at how AI can support the competency and how the human can support the competency,” Allen explains.

Part of the problem is that law schools lack a consensus on the right approach, and there is no universal blueprint that others can follow. In addition, policies on when AI should be introduced into the curriculum, what situations students can use it for, or whether or not it should be used at all vary widely depending on the professor and the school.

Likewise, the legal teams responsible for driving this change inside law firms, including those from professional development, innovation, and knowledge management, often are not working in sync. “Technology adoption tends to flow through innovation and knowledge management teams, while professional development teams own the actual learning strategy,” Allen explains, adding that these three groups need to come together to think about how to manage this new world.

A checklist for verifying AI legal work

Allen says she believes verification skills need to be explicitly taught rather than assumed. Firms and schools must intentionally show law students and practicing lawyers how to verify AI output, which is a critical skill in learning legal judgment.

To build this skill set, Allen borrows from a concept in the aviation industry that urges a comprehensive and repetitive approach. This process can build a verification checklist for law students and early career lawyers they can use repeatedly. More specifically, lawyers should:

• Start with source-level actions — Confirming that a traceable source exists for the AI output, that the source has been checked and confirmed (not just located), and that there is comprehension of the source beyond just the existence or verification status.
• Add a comprehension check and a citation check — This goes one step further by confirming the lawyer understands what that source says. This distinguishes verification from simple citation-checking because it is meant to test whether the lawyer could explain or defend the material independently.
• Tie the checklist to a written reflection artifact — The checklist can be formalized into a document, such as a judgment memo, that junior lawyers or students submit alongside their work. This would explain where they used AI, what they questioned, and how they identified any errors.

Once the checklist is developed, it needs to be baked into the first learning exercises of the education or training program from day one. To make it more effective, the checklist could be portable across a career with the same checklist a first-year law student uses evolving as they progress.

Listen to our recent Clarity podcast, How are law schools using AI and how will AI shape the future of legal practice? here

Likewise, it is important for the verification discipline to transcend from law school to first employer and be repeated until the checklist becomes part of lawyers’ DNA. Indeed, lawyers need to build the habit until it becomes second nature, at which point the explicit checklist fades into automatic behavior, while still being consciously reinforced periodically throughout their career.

Solving the junior lawyer "learning paradox"

AI is making it faster to produce polished work, Allen says, and this has resulted in some partners and senior associates quietly bypassing junior lawyers altogether. Unfortunately, that means the efficiency gains law firms want risk starving junior lawyers of the repetitions they need to build judgment in the first place. This learning paradox, she warns, will create huge downstream negative impacts in the not-too-distant future.

Effective and consistent mentorship and supervision are critical to avoid this impact and keep current law students and early career lawyers from falling prey to this learning paradox. Equally important is for those senior lawyers who supervise and mentor junior lawyers to be trained on how to evaluate AI-assisted work by reinforcing verification skills and the ability for junior attorneys to defend AI-assisted legal work independently.

Looking ahead, Allen sees simulation-based learning as the clearest path forward, and one of the biggest growth areas in legal training. In fact, Barbri itself plans to launch a simulation tool aimed at bridging the gap between law school and legal practice.

Simulations can offer a safe space to fail, learn, move on, and repeat, which allows lawyers to practice verification and reasoning without real-world consequences. Some law schools are already experimenting with this, and Allen points to examples of faculty innovation as a sign that schools are perhaps more advanced in this area than people think.

Still, Allen contends that technology alone will not solve the judgment problem and, in fact, may hinder it without proactive measures to develop judgment among junior lawyers. Clearly, AI fluency has to develop alongside human mentorship, relationship building, and legal reasoning, she says, adding that judgment is not happening by osmosis — it requires friction, failure, and guidance that, for now, only humans can provide.

You can register for the upcoming webinar, AI 101 & Why AI Creates Ethical Dilemmas for Lawyers, here