Getting Real About Lawyers and AI Tech?

Data and Trust,

They’re Problems.

The team asked me to share some thoughts on AI in the legal world. Makes sense—we now have a tech company delivering legal services as software to solve big, important problems and the momentum of recent tech adoption is undeniable.

In 2024, investors poured nearly $5 billion into legaltech — a 47% jump over the previous year, per Law360 analysis. According to the ABA’s 2024 Legal Tech Report, 30% of law firms surveyed say they’re using AI-based tools now, up from just 11% in 2023. Sure, some of that might just be legacy platforms like iManage or Relativity adding AI features—but still. Either way, it’s a huge shift.

With so many voices already in the mix, I had to ask: What’s left to say? But then I thought, let’s get real.

For AI to actually work, it needs good data to learn from. That means smart human prompts and expert human judgment to catch subtle, substance-driven nuances in output that require change. Let’s call that the Data Problem.

Also, lawyers are some of the hardest people to sell on tech. Especially when it “threatens” the work we’ve spent years mastering. Call it ego, pride, fear of being replaced, anxiety about ethics or quality—whatever the reason, resistance runs deep. Many lawyers (hmmm, most lawyers?) are wired to poke holes, find risk, and slow things down. Not the dream audience for innovation. Let’s call that the Trust Problem.

Combining the problems gets you low tech adoption and limited usefulness. Wah-wah. “I don’t do math—that’s why I became a lawyer.” “That’s not a legal issue, I defer to the business.” “It depends.” … said almost every lawyer, everywhere.

So, why do we have a Data Problem?

Most lawyers didn’t come up through math or science. Numbers, models, systems thinking—it’s just not how lawyers are wired. So when the conversation shifts to training data, model function, or learning iterations, eyes glaze over. We can crank out a merger agreement. But a clear, structured set of prompts to get AI to flag the right risks and, critically, spot the opportunities, apply legal tests in context, and actually mimic good judgment? Totally different skills, much more algorithmic thinking.

Reviewing, correcting, and iterating on AI outputs? Yeah, good luck getting a billing lawyer to block time for that. I’ve heard numerous times now from lawyers in the market, totally straight-faced: “I don’t understand, why can’t AI just do it?”

Also, not all lawyers are made the same. Just having a JD doesn’t mean someone has the substance or judgment to meaningfully train a model or use tech well within a structured delivery of a legal service.

Meanwhile, the lawyers most comfortable with tech—usually younger, raised on iPhones and Insta—are often still early in their careers. They don’t have the deep lawyering skills and experience that makes tech work really well and improve. It’s a real gap. Between what the tools need, and what most lawyers (right now) are equipped to give.

‘Okay lawyers, gather around, time for tech trust falls? Who’s going first…?’

Now, let’s talk about the Trust Problem.

Lawyers are trained to be skeptical. They also fear malpractice. Lawyers have practical concerns about data accuracy, bias, and transparency. They protect others against flawed non-lawyer (tech, human) legal reasoning. Attorney- client privilege. Confidentiality concerns. (The list continues.) Adding salt to the wound, stories abound of lawyers sanctioned or chastised for citing AI hallucinations in legal filings.

Oh, and lawyers generally mistrust anything made to help lawyers work that comes from non-lawyers or people claiming they know things just because they have a law degree.

Then, you’ve got the client- side considerations: How do you explain to a client that their $800+ per-hour attorney is using the same tool that the client is using for travel itineraries and to figure out what foodie haunt is next on their list? The optics can be problematic, even when the underlying technology is sophisticated and can be really helpful in improving outcomes, turn arounds and costs if used right and well.

Some of this mistrust, unfortunately, is pretty well- founded. For example, too many tech folks push products without any real understanding of lawyer liability, how lawyers practice, and what their day-to-day, unspoken internal lawyer hangups are.

Some of the mistrust is not. But let’s see you persuade the lawyer one way or the other.

So, yeah, trust is an issue.

We’re Doomed! Nah, there’s definitely an exciting path forward.

Alright, here’s the thing: both the Data Problem and Trust Problem are actually solvable. Easy? Not so much.

We need a total reset. We need a movement.

Tech makers—stop trying to replace legal judgment (or marketing like you do). You can’t; despite the hype, those AI models don’t exist. Let’s also quit calling everything else simple (like legal work doesn’t require knowing a boatload about context, legal analyses, human motivations and objectives, executive functioning, etc.) and treating all lawyers’ knowledge and skills the same. Both are just wrong.

If you want change, build tech that works with lawyers, not around or in spite of them. That means deeply understanding legal work: doing it, observing it, and collaborating closely with legal communities from day one through every test, iteration, and launch. (Credit where it’s due, some incumbents got where they are because they did exactly this.) You’ll get better data and this breeds trust.

Law firms—no more bans on new tools. Instead, try things like running lawyer-led tech hackathons. Challenge teams to find tools and reimagine workflows that reflect how they actually work, start to finish, while hitting specific client goals. (You know our more innovative counterparts in-house are already doing it!)

Yes, new tech + human solutions may impact the billable hour. But lawyers are sharp and resourceful. There are smart ways to price this work and reward those driving innovation. (Now, how to train young lawyers in this new world of tech + human, that’s a very interesting question for another article.)

The key is to accept that change is already here. So let’s lead it. When we redesign the work ourselves, we get to shape the inputs and outputs, choose who’s involved, and define how we engage with clients. So let’s trust ourselves (even if trusting outsiders is a bit difficult).

Law schools face some of the biggest challenges. Historically, they’ve been light on teaching practicum, though some now have strong experiential programs. But that’s not enough for what’s coming next (um, now).

What’s needed? Full immersion. We’re talking actual technology classes in law school, assignments and tests that require students to show substance and tech fluency, curriculum on non-law firm career paths, exposing students to legal work types and processes in-house and in product development. (First year lawyer requirements in law firms inevitably will change too, fyi.) Thankfully, we know of a select few amazing educators already heading down these roads.

(Shameless plug: We’ve kept all this in the front of mind as we’ve developed our jayaram.xyz tech portfolio. But seriously, we have because we believe in the movement. Lawyers can lead the way on tech.)

So, those are a few thoughts on legal tech. We’ve collectively got this! Solving the Data Problem and Trust Problem will be hard sledding (my fellow Canadians will get that one). But we have the grit, the literal duty to our clients, and—let’s be honest—we lawyers love control.

So let’s take the lead.

Join the movement.

The post Getting Real About Lawyers and AI Tech? first appeared on Jayaram Law.