TL;DR
- Legal AI adoption is now widespread, but trust hasn’t caught up: only about 22% of legal AI users report high trust in the output, and just 23% of in-house lawyers use AI daily (Vaquill AI / 2026 GenAI in Legal Benchmarking Report via Legaltech News; Bloomberg Law State of Practice 2026).
- Clients are raising the bar faster than firms can meet it: 70% of corporate legal departments expect AI-driven productivity gains from outside counsel, but only 3–6% believe most firms are actually delivering on quality, innovation, or productivity.
- Nearly two-thirds of firms haven’t changed their pricing model despite the efficiency AI is supposed to create.
- The trust gap has a measurable payoff: teams with high confidence in a tool report positive returns 89.5% of the time, versus 27.8% for teams that don’t trust it.
- Bottom line: the firms pulling ahead aren’t the ones with the most AI tools. They’re the ones that closed the trust gap first.
Surveys this year are saying the same thing: AI adoption in professional services is climbing fast. What those same surveys quietly reveal is that trust in the output isn’t climbing anywhere near as fast, and that gap is where most firms are losing ground.
The adoption numbers look good. The trust numbers don’t.
Roughly four in ten professionals now say their firms have embraced AI tools, twice the rate from a year earlier. But only 22% of legal AI users report high trust in the output, and just 23% of in-house lawyers use AI daily, while 27% haven’t touched it in six months. Wide adoption sitting on top of thin trust is the clearest signal in this year’s data.reuters+1
Why the gap exists
Three patterns explain most of it:
- Hallucinated citations have moved from demos to real filings. Even purpose-built legal research tools hallucinate 17% to 33% of the time, according to a peer-reviewed Stanford RegLab study, and practitioners now describe opposing counsel filing motions built on cases that don’t exist.
- Some tools add a verification step instead of removing one. A widely discussed complaint in legal and accounting communities: upload the document, read the AI summary, then read the whole thing anyway to check it. If verification takes as long as doing the work, the tool doesn’t earn a place in the workflow, and professionals quietly stop using it within two weeks.
- No one can trace who made the call. When an AI-assisted output turns out wrong, firms without clear logs of who reviewed and approved it can’t answer a basic question: who was accountable for this decision. That ambiguity is a bigger blocker to adoption than skepticism about the technology itself.
What clients are actually asking for
Corporate legal departments have stopped treating this as optional. Nearly a third say they’re reconsidering relationships with outside counsel who can’t show clear AI-enabled value within 12 months, and almost a quarter of law firm professionals say they’d refuse a job offer from a firm without professional-grade AI tools. At the same time, more than a third of professionals admit to using unauthorized “shadow AI” at work, which creates exactly the security and compliance exposure firms are trying to avoid.
What closes the trust gap
The firms getting a real return aren’t buying more tools. They’re doing three things in order: grounding AI use in sources the team already trusts, making review and sign-off traceable so accountability is never ambiguous, and training staff on what the tool is reliable for versus what still needs a human check. That’s a workflow decision, not a procurement decision.
Bottom line: adoption without trust is a liability, not a win. The gap between “we use AI” and “we trust what it produces” is exactly where governance, training, and clear ownership earn their value.
Sources: Thomson Reuters Institute, “2026 Future of Professionals Report”, Vaquill AI, “What Lawyers Really Think of Legal AI in 2026”, Reuters Breakingviewsreuters+2
Aspen Management Group works with boutique advisory firms to clarify key workflows, improve efficiency, layer in AI where it adds value, and build governance and training around that change
Scott spent 20 years running a managed IT services practice with law firm clients across the DC Metro area, and has worked in technology for 30 years. AMG helps boutique law firms get practical value out of AI.