The next AI divide in law may be between lawyers who use artificial intelligence (most, these days) and lawyers who know how to think alongside it, said Catherine “Cat” Casey, founder of TechnoCat.
For two decades, Casey has told lawyers that AI technology would revolutionize how they practice law. For years, the response was often: “Enough already,” she recalled.
That has obviously changed, with AI embedded in systems surrounding legal work. Yet Casey — who began her career at the intersection of law and technology, including through roles in eDiscovery and innovation at Gibson Dunn, PwC and KPMG — remains concerned, noting, “The more consequential question is whether lawyers will develop the judgment and fluency necessary to work effectively alongside increasingly capable machines.”
Understanding the Tool
Some of the most publicized AI failures follow a predictable pattern: A lawyer taps generative AI, accepts its output and discovers too late that the system produced fabricated information. The lesson, Casey said, “shouldn’t be that AI is inherently untrustworthy.” Instead, she encouraged lawyers to develop workflows that incorporate verification “because hallucinations aren’t going anywhere,” she said. “It’s a feature in how generative AI works.”
She compared partnering with AI to working with attorneys just out of law school: “You would check their citations. Why are you trusting a machine?”
AI Washing and Real Choice
Casey also worries about “AI washing,” or attaching AI terminology to products without clearly explaining what the technology does. “Everyone’s claiming they have the same flavor of AI, which is making it hard for people to choose,” she said.
She encouraged law firms to select technologies by identifying the problem the firm is trying to solve, then consider how a technology could fit into existing workflows and what risks it would introduce.
Casey is especially mindful of the lawyers and staff who make up a firm’s “bread and butter” workforce. “Everyone needs to be trained on how to understand the risks and opportunities broadly of AI, not just how to use a tool,” she said.
Start with a Refrigerator, Not Client Files
For lawyers overwhelmed by the AI landscape, Casey recommended starting somewhere less consequential than a legal matter. “Much like riding a bike, start doing low-risk, high-frequency things with AI,” she said. “Use AI to plan a road trip. Take a picture of what’s in the refrigerator and ask for a recipe. The point is repetition,” as generative AI requires iteration, or asking, refining and correcting.
Casey called the resulting skill “portable AI fluency,” or an understanding of how to work with these systems that can transfer from one tool to another.
Colorado AI Regulations
For Colorado lawyers, the state’s AI regulatory framework provides another reason to examine AI’s impact on people and law work.
In May, Gov. Jared Polis signed Senate Bill 26-189, the Automated Decision-Making Technology Act, which repealed and reenacted provisions of the 2024 law with new requirements for automated decision-making technology used in consequential decisions.
Casey said regulations should remind attorneys to look at how AI is being deployed. She continued, “What are the uses that it is being deployed for and how does it impact humans? What happens when it gets something wrong? And where is the human who can intervene?”
The Shift in Legal Work
Casey called the current AI moment “not an existential but an evolutionary crisis” for the legal profession.
She used the ATM as an analogy for the AI revolution. Rather than eliminate banking or bank employees, the technology changed economics and expanded what banks could do.
Legal AI is producing a similar result, she said. The technology enables attorneys to handle more matters, serve clients who previously couldn’t afford legal services or develop new areas of practice. “Look instead at what doors are now open,” she said.
Lawyers have an advantage that has little to do with technology, too: language. Generative AI is largely controlled through language, and lawyers spend their careers thinking about precision, ambiguity and the consequences of a single phrase. “Syntax for a lawyer is a superpower,” she said.
Lawyers will require enough substantive knowledge to recognize when an AI-generated answer doesn’t make sense and enough command of language to direct it effectively.
Casey joked, “There probably won’t be a day when lawyers arrive at work and discover robots have taken every job. But some lawyers may become dramatically more effective than others because they know how to use AI.”
For the legal profession, that may be the more important AI divide: the line between lawyers who know how to use the machines effectively versus those who don’t.