Why the best lawyers are learning to talk to AI

Industry leaders have been blunt. Lawyers need to wake up to the reality of AI or risk becoming professionally obsolete. That warning is a bit dramatic because AI isn’t replacing legal professionals wholesale, but it is reshaping the value they bring and how fast they can deliver it. The lawyers making the biggest impact today are the ones who understand how to guide artificial intelligence with precision and control. Think walking by foot versus driving a vehicle – a new skill set and a new vocabulary.

I’ve been out of Big Law for over two years, which is long enough to notice where my own skills have frayed and where they’ve evolved. When an AI startup asked me to create their privacy compliance suite, I approached it with the kind of wariness lawyers are trained for. I was handed a ChatGPT-generated privacy policy filled with undefined terms and placeholders referencing documents that didn’t exist. In other words, a mess, but it was also a gold mine.

I uploaded the document into ChatGPT, and the interface shifted into something closer to collaborative drafting. Instead of starting from scratch, I prompted the model to revise line by line and adjusted how it framed risk. I fed in examples of clauses from comparable companies and asked for output tailored to those formats. When a provision was highly bespoke to the client’s unique offering, I stepped in to do the deeper legal work – but for the bulk of it, AI handled the drafting and research in hours.

Compare that to a friend who spent nearly $7,000 building out a similar privacy suite through traditional legal channels. Most of that cost came down to the slow pace of drafting and unclear requirements. I reached the outcome faster and with tighter alignment to the client’s product.

AI has changed the legal skills stack

There’s a popular tendency to describe AI as disruptive. That’s accurate, but incomplete. Just as the arrival of Excel didn’t eliminate accountants, AI won’t eliminate lawyers. It is, however, recalibrating what legal professionals need to know to remain effective.

In high-performing legal teams, we’re seeing the rise of a distinct AI-legal skill set that wasn’t part of traditional training. These aren’t technical skills in the traditional sense – you don’t need to know how to build a large language model (LLM), but you do need to know how to frame legal questions in a way that AI can respond to.

You need to know when to trust AI-generated output, and when to override it. You need to be able to translate legal logic into structured prompts that guide a model toward the right outcome. You need to evaluate tools for more than convenience, focusing instead on how they perform under legal scrutiny.

What’s more, these skills are starting to show up in hiring decisions. In-house teams want lawyers who can iterate faster. Clients are asking their outside counsel how they’re using AI to create value. Legal departments are rethinking workflows, using AI for early-stage drafting and research. Ultimately, AI is becoming a structural layer in how legal teams deliver insight.

Cross-examining algorithms

To be clear, talking to AI is more than typing into a chatbot. In legal practice, you’re using your own judgment to steer the model through ambiguity and shape its output into something that meets legal and commercial standards. That might involve asking a model to generate clause variations along a risk spectrum.

It might mean taking the first pass of a diligence memo and filtering it through three different lenses – legal risk, business impact, and jurisdictional nuance. Lawyers are used to cross-examining people, now we need to get comfortable cross-examining algorithms.

None of this replaces legal training, but it does challenge lawyers to apply that training differently. To be a great lawyer today, you still need to apply sharp judgment but with quickness and clarity.

The next generation

These skills are being learned in real time. What’s striking is that lawyers aren’t approaching this like a technology problem, but instead treating it as a professional evolution. They want to understand how to use AI, and how to supervise it. They’re learning how to frame legal tasks for AI systems with enough specificity to avoid hallucinations and enough flexibility to accommodate ambiguity.

There’s a cultural shift underway, too. Lawyers who once defaulted to AI hesitancy are starting to move toward hands-on experimentation. They’re building muscle memory around prompt design, testing outputs across multiple models and generally asking better questions. In doing so, they’re creating a new baseline for what legal excellence looks like.

AI hasn’t erased the need for legal judgment. If anything, it’s made that judgment more visible – because nothing AI produces carries weight until a lawyer decides how it holds up in the real world – at least for now.

We've featured the best AI writer.

This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro