Are Law Firms Training Attorneys Adequately for AI?

AI training is now common in law firms, but access, content and practical support remain uneven—especially in smaller firms. Primary Research Group found that 63% of surveyed firms offered some form of AI training: 28.57% provided mandatory training and 34.69% offered it on an optional basis. Nearly 29% offered none.

The firm-size divide was stark: 56.25% of firms with fewer than 23 lawyers had no program, while none of the firms with more than 1,000 lawyers reported having no training. Overall, 57.14% of attorneys had personally received formal AI training. That figure rose to 76.92% in the largest firms but was 50% in the smallest. The gap may become a competitive and risk-management issue as AI features move from experimental tools into research, drafting and document workflows.

Training content also reveals unfinished business. Ethics was covered for 43.88% of respondents, drafting or document review for 42.86%, and legal-research tools for 38.78%. Only 14.29% reported training focused on verifying the truth or accuracy of AI output—even though confident-sounding errors are among the technology's most important professional risks.

Firms are also missing a chance to make good practice reusable. Only 29.59% reported that their firms maintained an AI prompt library, and just 20.41% of attorneys had used one. Adoption reached 61.54% at the largest firms, suggesting that shared prompts and templates are becoming part of institutional knowledge management rather than merely an individual technique.

The result is a divided workforce: 39.79% felt prepared for the future of AI in legal practice, while 39.80% felt unprepared. Closing that gap requires more than a one-time ethics presentation. Firms need role-based practice, verification routines, approved tools, confidentiality rules and examples tied to real matters. Training should make safe behavior easier to repeat—not simply prove that a session occurred.