The legal sector is integrating AI at a pace that has far outrun its capacity to manage the psychological risks that come with it. That is the central finding of a new report by Positive Group, a behavioural science consultancy, which warns that law firms risk what it calls “cognitive surrender”: a creeping over-reliance on automated tools that erodes the professional judgment that legal practice depends on.

The report, titled The AI Leadership Challenge in Law, draws on qualitative research with 16 senior leaders across global commercial firms, including Managing Partners, Chief AI Officers and innovation leads from Orrick, Herbert Smith Freehills Kramer, Bird & Bird, Baker McKenzie, A&O Shearman, White & Case and Gilbert + Tobin. It was produced in collaboration with researchers from Harvard Business School, RSGI and Hubel Lab.

A market running ahead of its guardrails

The report lands against a backdrop of record capital flows into legal AI. Swedish platform Legora reached a $5.55 billion valuation following a $550 million Series D led by Accel, while rival Harvey is valued at $8 billion and was reported to be seeking to raise at $11 billion. Venture funding for legal tech startups hit a record in 2025, with the sector raising $4.08 billion, a 77 percent increase on 2024.

That capital influx has accelerated tool deployment across firms of all sizes. The Positive Group report argues the pressure to adopt has outpaced the psychological infrastructure needed to use AI responsibly.

The consequences are already visible in courtrooms. Trackers of AI hallucination cases in U.S. courts have catalogued more than 500 instances of courts commenting on or sanctioning AI-generated content in legal filings, with 203 cases recorded in 2026 alone and 279 in 2025. A broader global database compiled by lawyer and data scientist Damien Charlotin counts more than 1,300 cases worldwide. In one notable U.S. case, a retired magistrate judge admitted he had almost included AI-fabricated citations in a court order, writing that the plaintiff’s use of AI had “affirmatively misled” him.

From compliance to behaviour change

Positive Group’s central argument is that policy-level responses to these failures are inadequate. Issuing AI usage guidelines, the report contends, does not address the underlying cognitive mechanisms that make automation bias a persistent risk. The consultancy identifies junior lawyers as particularly exposed, given their heavier reliance on AI for routine workflows and the reduced oversight those tasks typically attract.

The recommended corrective is not a technology fix but a behavioural one: structured frameworks that define the precise level of human input required for specific task tiers, and active cultivation of the habit of scepticism at every stage of AI-assisted work.

Ben Allgrove, Partner and Chief Innovation Officer, Baker McKenzie

“We need to treat AI like a junior lawyer by checking it, questioning it, and training scepticism as a core skill. The risk is machine dependence, so we must balance speed with judgment.”

Skills that do not change

A recurring theme across the report’s senior contributors is that the core attributes of good legal practice remain unchanged by the tools now available.

Francesca Bennetts, Partner, A&O Shearman

“The skills that make a good lawyer, critical thinking, creativity and attention to detail, are the same skills needed to be an AI-augmented lawyer. The tools may have changed, but the foundations of good practice remain the same.”

Caryn Sandler, Partner and Chief Knowledge and Innovation Officer at Gilbert + Tobin, pointed to a shift in how firms are approaching junior development in response to these pressures.

Caryn Sandler, Partner and Chief Knowledge and Innovation Officer, Gilbert + Tobin

“It’s not just about training juniors in legal content, we’re also building mindset by interrogating information, questioning sources and not taking everything at face value. There’s now a focus on continuous improvement, empathy and leadership skills much earlier in their careers.”

Leadership polarisation as a compounding risk

Beyond individual lawyers, the report identifies fragmented leadership as an additional obstacle. Polarisation among firm leaders, ranging from uncritical enthusiasm for AI to outright resistance, is actively undermining ROI on technology investments and stalling coherent decision-making. The report calls for leaders to model calibrated AI use rather than occupying either extreme.

Positive Group’s AI Leadership Challenge in Law is available to download from its website. The research forms part of a broader cross-sector programme on AI leadership and organisational maturity, with findings recently referenced in Harvard Business Review.