Despite common misconceptions, AI isn't here to replace lawyers and legal teams; it's here to help them. Rest assured that humans retain all the power when it comes to legal applications of AI—as they should.
Whether you need to overcome a lingering sense of fear for yourself, or help a colleague or client through it, our 2025 AI Visionaries are here to help.
Specifically, they impress three key reasons each of us should start to feel a sense of growing confidence in an AI-powered legal career.
#1: AI won’t just help legal teams work faster or cheaper; it’ll help them work better.
Yes, it’s great that AI can help you accelerate timelines and deliver results using a smaller chunk of change.
But more important than that is AI’s ability to improve the quality and accuracy of your work, especially on traditionally exhausting, human-error-prone projects like document review.
“For me, as a busy mother of two, AI helps get the job done right the first time, aligning with my focus on effective and clear communication. It’s not just about speed; it’s about achieving better results, and defensibly,” says Fiona Campbell, head of electronic disclosure at Fieldfisher LLP.
“Lawyers, bound by practicing certificates and their duties as officers of the court, must stand over their work, using AI as a supportive tool rather than a replacement,” she continues. “By streamlining labor-intensive tasks and providing robust audit trails, AI enables better, faster outcomes for both teams and clients.”
“Lawyers, bound by practicing certificates and their duties as officers of the court, must stand over their work, using AI as a supportive tool rather than a replacement,” she continues. “By streamlining labor-intensive tasks and providing robust audit trails, AI enables better, faster outcomes for both teams and clients.”
Fiona Campbell, head of electronic disclosure at Fieldfisher LLP
Legal practitioners should feel eager to adopt AI, therefore, knowing that they won’t just be chewing through projects faster—they’ll be providing their clients with better service, too.
“The prize we have been chasing in e-discovery for the last two decades is how to do review cheaper and faster, but also with better quality results. Generative AI is a tool that potentially can move us closer to that previously unachievable goal,” says Matt Jackson, counsel of data analytics and e-discovery at Sidley Austin LLP.
Of course, our AI Visionaries note, too, that part of the assurance they feel about the quality and pace of their results with AI comes from the careful vetting and validation that are essential to any legal technology implementation.
“People have sorted into one of two camps: one that claimed generative AI would make life unrecognizable within a matter of months, and the other that was ready to dismiss it as a useless distraction from the very start. It always seemed to me that the answer was somewhere in between,” explains Matt Kelly, counsel at Debevoise & Plimpton LLP.
“The technology held incredible promise based on its ability to enhance access to information, to improve operational efficiencies, and to create interactions that felt uniquely meaningful to users, but it also seemed clear that it would take time for that promise to become reality,” he continues.
Both technical innovation and human adaptation to working with these tools are essential in helping AI deliver on its potential.
#2: Artificial intelligence can’t work in legal without human intelligence to guide it.
Make no mistake: LLMs simply can't mimic human problem-solving, ethics, and creativity. The practice of law is one where human insight is not just important—it’s essential.
“The idea that lawyers will be easily replaced by AI systems simply because AI systems are effective at knowledge retrieval fundamentally misunderstands the service that lawyers provide. Because knowing the answer to a client’s question is often the easiest part of the job,” Matt Kelly emphasizes for us. “Knowing what to do with that answer, and how to provide advice based on the answer in a way that adds value for your client? That is the real work.”
“The idea that lawyers will be easily replaced by AI systems simply because AI systems are effective at knowledge retrieval fundamentally misunderstands the service that lawyers provide. Because knowing the answer to a client’s question is often the easiest part of the job. Knowing what to do with that answer, and how to provide advice based on the answer in a way that adds value for your client? That is the real work.”
Matt Kelly, counsel at Debevoise & Plimpton LLP
Your expertise is needed to make the best use of AI in your projects, and your clients are looking to you for the same sage advice they’ve always needed.
“Law relies on human judgment, creativity, and ethics, which are unique qualities AI cannot replicate. Instead, AI complements lawyers by handling data-heavy, repetitive tasks, freeing them to focus on advisory and strategic roles,” Fiona notes. “This partnership enhances, rather than replaces, the human role in law.”
“e-Discovery has a 2 percent problem: only a very small number of documents actually matter in litigation and investigations, but the time and expense to get to them are enormous and disproportional,” Matt Jackson adds. “The generative AI tools I have worked with are very efficient at finding the needles in the haystack that actually matter, which is incredibly valuable.”
And from there, the really impactful work begins: the strategizing, advising, collaborating, and storytelling that the human mind does best.
#3: An insanely talented legal community is ready to help you understand and implement AI.
When technological innovation happens so fast and discussion topics around AI evolve even faster, it’s easy to feel like you’ve fallen behind. But fear not: even if “agentic AI” still doesn’t mean much to you, there's still time to learn about AI and how to start using it to forge a forward-thinking career path.
Engaging with the community is an essential and particularly useful way to educate yourself on AI applications for the legal space, as well as best practices for implementing it in your day-to-day work.
“Community is vital in this space, as collaboration fosters shared knowledge, builds trust, and ensures that AI is developed and applied ethically and effectively across the profession,” Fiona says.
Attend events like Relativity Fest and ILTACON to get your feet wet. There, you can get hands on with AI and discuss its more philosophical and theoretical facets with industry peers who are interested in the topic, too.
Additionally, make sure you look to your partners—technology and service providers—to help you navigate this learning process. They shouldn’t just be ready and willing to help educate you on AI and what they can and can’t offer around it; they should be eager to explore the topic with you, teach where they can, and gather your feedback to inform their own strategies.
“It is really important to have a trust-but-verify approach to using gen AI (or any technology). Challenge the results; be intellectually curious,” Matt Jackson advises. “Ask questions about the underlying LLM: How often is it updated? What safeguards are built into the tool to prevent or detect hallucinations? What metric driven testing has the provider done to measure the efficacy of the product? Why seemingly minor tweaks to prompt criteria generate significantly different results?”
“It is really important to have a trust-but-verify approach to using gen AI (or any technology). Challenge the results; be intellectually curious. Ask questions about the underlying LLM: How often is it updated? What safeguards are built into the tool to prevent or detect hallucinations? What metric driven testing has the provider done to measure the efficacy of the product? Why seemingly minor tweaks to prompt criteria generate significantly different results?”
Matt Jackson, counsel of data analytics and e-discovery at Sidley Austin LLP
I bet you’ll be surprised and delighted by how quickly you’ll pick up on this subject and build confidence in it. There’s just so much to explore, and so many people ready to chat about it with you.
And once you feel that confidence? Give back! Participate in the conversation as an informed advisor. Be ready to help others get comfortable—especially clients who may have questions about the tech and what makes it worthy of their trust.
Matt Kelly has great advice for working with clients who may feel that hesitation.
“I tend not to push too hard. I will generally point to industry or peer adoption, if relevant under the circumstances. And if I can point to our own track record of success with the specific product or technique, then I will do that,” he says. “But that is about it. I never want my clients to think I’m more interested in the technology than their preferences. And so I think it is very important for us—particularly as AI-supportive practitioners—to be as flexible as they want us to be.”
Graphics for this article were created by Natalie Andrews.
