Inside the growing judgement gap

As law firms and in-house legal teams adopt AI for research, many are simultaneously cutting back on the senior associates and paralegals who provide the due diligence. This removes a critical layer of review, leaving firms exposed to harmful errors. We are building law firms that can't catch their own mistakes, and the people who could have prevented this are already gone.

I call this the Judgement Gap, and it's already a live issue in UK legal services. It is no longer hypothetical, and it is no longer rare. There are now 64 documented UK cases involving confirmed or suspected AI hallucination in court filings; fabricated citations, miscited authorities or references to cases that simply do not exist, and that number is growing monthly. Judges in England and Wales have raised concerns about AI-assisted legal filings and hallucinated citations, and the Law Society has warned solicitors to verify AI-generated legal research and citations rather than just relying on them. 

Knowing how to use AI tools is not as important as having the experience and seniority to know when not to trust them. With organisations getting smaller as these people leave, this judgement disappears with them.

What the Departure of Experienced Workers Means for Law Firms

There’s a pattern that keeps appearing. A firm invests heavily in AI. It feeds years of institutional knowledge into the system, such as processes, decision frameworks, historical data, and customer intelligence. The firm then gives this tool to its workforce and tells them to use it. The problem? The people who would have known whether the AI's output was right or wrong have often already left. Some have retired, some were made redundant - often to fund the very AI programme that now operates without oversight.

This is not a hypothetical situation. It's happening right now across the UK in the law industry, and almost nobody is talking about it with the urgency it deserves.

For years, the departure of experienced employees has been treated as an unfortunate but manageable event. Through knowledge management programmes, documentation initiatives and mentoring schemes, firms have built entire functions trying to capture senior knowledge before people leave. But these programmes capture the information, not the judgement. What is lost was never just the facts someone held. It was their ability to look at a recommendation, a data pattern, or a proposed course of action and say, "that doesn't feel right" and be correct based on twenty years of experience.

I call this judgement the Verification Layer, and we are systematically dismantling it at the moment we need it the most.

The Importance of Expert Verification

The timing could not be worse. AI adoption across legal services is accelerating, and within such a time-pressured industry, it offers genuine efficiency gains. Research I've been involved in shows that whilst around two thirds of leaders and employees believe they can keep pace with AI, over 90% of organisations report that workforce barriers are limiting their progress. 

Whilst people believe they are keeping up, the firm’s ability to verify whether AI is giving them good answers is quietly eroding beneath them. 

Consider the scenario now playing out in firms across the country. A newer employee, who is hardworking, keen to prove themselves, and digitally fluent, receives an AI-generated piece of analysis. They've been told this tool has been trained on the firm’s historical data and institutional knowledge and have been encouraged to use it. They do not have over twenty years of domain experience to tell them the output has made an assumption drawn from a data set that was flawed when it was created five years ago. The person who understood that limitation has since left. So, the new employee acts on the output. Multiply this across an entire UK law firm and you have a systemic verification failure hidden behind a veneer of AI-enabled ‘efficiency’. I’d describe this as a Judgement Gap, and it is widening with every experienced person who leaves. 

And here is the part most firms miss: this is not a failure of diligence. Research published in Harvard Business Review in May 2026 found that AI tools create what researchers call ‘psychological debt’; a cluster of effects that includes cognitive offloading, where the brain stops exercising the critical-thinking skills that AI now handles. The more an employee relies on AI-generated output, the less capable they become of spotting when it is wrong. So, the Judgement Gap is not only about losing experienced people. It is about what happens to the thinking capacity of the people who remain.

Where the Judgement Gap is already breaking down

The courts are providing a steady stream of cautionary examples. In Ayinde v London Borough of Haringey [2025], a barrister and solicitor submitted five fabricated cases to the High Court in a judicial review, with one purporting to be from the Court of Appeal. When the court asked them to produce the authorities, they could not, because the cases did not exist. In her subsequent ruling, Dame Victoria Sharp warned that generative AI tools ‘can produce apparently coherent and plausible responses’ that may turn out to be ‘entirely incorrect.’

More recently, a High Court judgment saw a top-20 UK firm criticised by the court after a junior associate cited a fabricated statutory provision. The judge described the failure to check the legislation as ‘inexcusable,’ noting there was no evidence the associate had checked the AI-generated references. They had, in effect, relied on the AI to do the thinking. The firm self-referred to the Solicitors Regulation Authority. It was the third such public reprimand in a matter of months.

These are not isolated lapses by careless individuals. They are the predictable result of a structural problem: AI capability is being deployed faster than the human capability to verify it, and in many firms the people who could provide that verification have already been shown the door.

An Impending Compliance Crisis

In regulated industries, the Judgement Gap is heading towards a compliance crisis. And the regulatory response is already underway. In February 2026, the Civil Justice Council launched a formal consultation on whether new rules are needed to govern the use of AI by legal representatives in preparing court documents. Notably, the CJC has proposed that solicitors and barristers should be required to confirm they did not use AI to generate the content of witness statements intended for trial - a direct response to exactly the verification failure described above.

The direction of travel is clear. Increasingly, regulators will ask not just whether a firm uses AI, but whether the human expertise to challenge it still exists within the organisation. Firms that pre-emptively build Verification Roles - positions with the explicit job of interrogating AI outputs using genuine domain expertise - will find themselves ahead of the curve. This is not a theoretical capability-planning exercise. It is a risk-management imperative.

Three Things Law Firms Can Do to Avoid a Judgement Gap 

First, firms must reframe the situation. They need to stop treating AI as a replacement for departed expertise and start treating it as a tool that makes existing knowledge more powerful. AI can hold the information, but humans must hold the judgement. The two are not interchangeable and pretending they are is a fast route to looking efficient whilst exposing clients to serious risk. 

Secondly, invest in critical thinking as a core capability, not a nice-to-have. This requires far more than a training module. It means encouraging employees to question AI’s output and pairing less experienced employees with those who still have the domain depth to model what good questioning looks like. It also means accepting that this takes time and costs money - but far less than the cost of a negligence claim arising from an AI output that nobody had the expertise to challenge.

Finally, rethink what experience means in your hiring process. If your entire workforce is optimised for AI use, but nobody has deep domain expertise, you've built a team that can operate the technology, but can't determine if it's heading in the right direction. The firms that retain or recruit for judgement specifically, not just capability, will have a decisive and durable competitive edge.

The Capability your Competitors Cannot Automate

Critical thinking has always been valuable. In the age of AI, it becomes the thing your competitors cannot automate, your regulators will demand, and your clients will judge you by. With high-profile cases showing the risks at play when AI errors reach the courtroom, the Judgement Gap is becoming an increasingly urgent problem rather than a distant one. 

The question is not whether your team can use AI. It's whether they know when not to trust it. In too many UK law firms right now, the honest answer is that the people who could have caught the mistake have already left. The firms that treat experienced judgement as a cost to be cut will discover that it was the most valuable capability they had.