Your Lawyers Are Already Using AI. Almost Nobody Trained Them.

Individual lawyers have embraced AI faster than almost anyone predicted. Adoption more than doubled in a single year, and the tools have moved well beyond novelty use. Lawyers now regularly use generative AI for drafting correspondence, general research, brainstorming, and summarizing documents, tasks that make up a meaningful share of daily legal work.

At the firm level, the picture looks completely different. More than half of firms haven't provided any training on responsible AI use, and have no plans to. That's not a small oversight. It means a majority of firms have lawyers actively using AI on client matters with no firm-level guidance on what's safe, what's not, or what good practice even looks like, the same strategy gap between usage and governance showing up as a training gap too.

It's tempting to think the risk here is AI itself, hallucinated citations, data exposure, inaccurate output. Those are real risks, covered elsewhere. But the deeper issue is that individual adoption without training means every lawyer is independently deciding what's acceptable, with no consistency across the firm.

Professional obligations around competence, confidentiality, and supervision don't pause because the technology is new. When lawyers experiment with AI tools on their own, without firm-wide standards, the risk isn't just that one person makes a mistake, it's that the firm has no way of knowing how many people are doing what, with which tools, on which types of matters.

Firms aren't avoiding this without reason. Concerns about data security, ethical compliance, privilege protection, and trust in AI-generated output are legitimate barriers, not excuses. The instinct to be cautious is reasonable.

The problem is that caution without action doesn't actually reduce risk, it just delays addressing it while individual usage keeps climbing anyway. Waiting for AI to feel fully settled before training anyone means training arrives after the exposure has already built up, not before, including the kind of ungoverned AI usage that raises cyber risk.

The training gap isn't just an ethics or malpractice exposure question, it's increasingly a talent and competitiveness issue too. A meaningful share of legal professionals report considering leaving their firm within two years specifically due to a perceived gap between what they need to do their job well and what the firm actually provides. Separately, inadequate training is now identified as one of the most persistent barriers to firms actually realizing value from their AI investment, not lack of tools, lack of guidance on using them.

Firms spending on AI tools while skipping training on how to use them are, in effect, paying for adoption without paying for the part that makes adoption safe and effective.

This doesn't need to be an elaborate curriculum. A workable starting point typically includes:

• A short, mandatory session covering which tools are approved, what data can and can't go into them, and why
• Clear examples of what good AI-assisted work looks like versus what creates risk, using real scenarios relevant to the firm's practice areas
• A named point of contact for AI-related questions, so uncertainty doesn't just get resolved by individual guesswork
• A plan to revisit and update training as tools and firm usage evolve, rather than treating it as a one-time event

The defining risk in legal AI right now isn't the technology outpacing what firms can handle, it's firms letting adoption outpace guidance. Closing that gap is one of the most actionable, lowest cost improvements a firm can make, and it directly addresses the training barrier that's already holding back the value firms are trying to get from AI in the first place.

Not sure where your firm's current AI usage stands, or what training gaps exist across your team? Our AI Tools Assessment reviews your workflows and flags exactly where guidance is missing.