Scaling up: The AI-dependent lawyer
Lawyers no longer need to be convinced to use AI. They need to be able to trust it, and that starts with the legal sources behind it.
The legal community’s rapid adoption of AI should not be mistaken for unquestioning confidence in it.
Regular use has done little to ease concerns about inaccurate outputs. If anything, those concerns are rising, alongside a growing preference for AI grounded in trustworthy legal sources.
In this report, we look at how lawyers are scaling AI use, where it is becoming embedded, and why trust matters more than ever.
Part 1
Trust is the metric on everyone's mind
We've been collecting data on generative AI adoption in the legal community for three years now, right when it was first being introduced to the market. Our latest survey of 543 UK legal professionals shows adoption has risen to 94%.
Despite the optimistic uptake, we've also recorded an unwavering demand for the quality and authority of the information powering these tools.
The question increasingly shaping AI adoption is: can I trust the output?
Our mid-2026 survey found four-fifths (81%) of lawyers feel more comfortable using AI that is grounded in legal sources, a noticeable increase from 70% a year ago, and from 65% at the beginning of 2024.
This is especially true for lawyers working in-house (88%) or across regulatory & compliance (85%), private client (84%), commercial & corporate (83%), and employment (83%) roles.
We have already seen examples in the market where AI mistakes attract significant attention, says Bhavisa Patel, Director of Technology Products at Eversheds Sutherland.
"No firm wants to be known for using AI badly. Once confidence is damaged, it can be very difficult to rebuild."
This largely comes down to the source material, says Patel.
"Lawyers need to be able to verify sources and test outputs against underlying material. AI-generated work should not be treated as reliable simply because it looks polished."
Concerns over inaccurate AI outputs continue to rise
Fabricated or inaccurate information has been the top concern since we first started measuring it at the beginning of 2024.
Yet the percentage of lawyers concerned about AI outputs has risen from 57% to 83%.
This indicates that, while most AI tools have introduced features that link to source material, it is the quality of the underlying source material that is in question. It also suggests that using AI tools has done little to alleviate concerns.
There are some slight variations in the data. Respondents from the Bar are slightly more concerned about hallucinations (87%) than the wider group, while senior leaders expressed more concern than mid to junior level employees (86% vs 82%).
Alongside hallucinated or inaccurate legal content, many expressed concern about confidential data leakage (53%) and becoming too reliant on AI (52%).
Dr. Gerrit Beckhaus, a partner at Freshfields and Co-Head of Freshfields Lab, says AI should support legal judgement, not replace it.
"A qualified lawyer must remain accountable for the work product, with appropriate review built into the workflow."
Tony Randle, Director of Client Tech & Service Improvement at Shoosmiths, says firms remain accountable for the final work product, not the AI tools themselves.
"Lawyers using AI should be able to demonstrate where information came from, how outputs were reasoned and verified, and what controls are in place to protect client data."
He continues: "If you cannot retrace all the steps from source to answer, you have not got governance."
And the stakes are high. Randle says mistakes involving AI won't simply require an apology to the client.
"They could involve negligence claims, a regulatory breach, court sanctions and reputational embarrassment."
Perhaps most importantly, organisations should have a backup plan in place says Peter Lee, Partner and Global Head of AI Governance at Simmons & Simmons.
"They should ask what could go wrong if the AI produces an incorrect result and whether they have the governance framework to manage that risk."
AI is causing more clients or the general public to misinterpret the law
More than three-quarters (77%) of respondents told us AI is causing clients or the general public to misinterpret the law.
A mid-level associate at a large law firm says this could be a result of the AI tools they're using.
“Clients only have access to general AI tools at the moment, not the tools with the right harness that they can actually rely on.”
Concern is consistent across segments, from 72% in medium-sized law firms to 80% at the Bar, with 83% among academics. This is particularly the case for criminal lawyers, although the base was small.
A leader at a large law firm told us they've noticed a "huge increase in litigation, as claimants are able to harness AI to bring claims, not all of which will be meritorious."
Lawyers care more about security and transparency than any fancy feature
The importance the legal sector places on trust is also reflected in the criteria lawyers use when selecting a new technology provider. The three leading considerations are strong data privacy and security standards (59%), transparency about how AI is used (40%) and human oversight of AI-enabled services (39%), all ranking ahead of cost and commercial value.
Unsurprisingly, lawyers working in regulatory & compliance (70%), litigation (67%) and employment (65%) placed particular importance on strong data privacy and security standards.
Randle says law firms and legal teams need "evidence that goes beyond mere vendor assertions during the sales process."
If challenged by a client or regulator, the process should be fully explainable and auditable, he says.
Lee advocates organisations being open about where AI is being used, what role it plays in decision-making, and where human review remains critical.
"People are generally more willing to embrace AI when they understand its purpose, its limitations and the safeguards surrounding its use."
A general counsel that oversees a large legal team told us: “Trust is critical to building reputation and remaining relevant in business. AI must support, rather than undermine, that trust.”
They add that ineffective governance can put both risk management and reputation in jeopardy, particularly where AI use leads to failures in complying with laws and regulations.
As an organisation it is critical you have a holistic understanding of what you're buying, says Patel.
"You need to properly understand the functionality, the data flows, what the tool can access, how outputs are generated, and how people are likely to use those outputs in practice."
Existing supervision, information security and data governance processes will need to evolve so that AI-generated output is not only checked by the person using the tool, but reviewed through the right supervisory and risk framework, she says.
"One of the challenges with AI is that the bigger risks are often not obvious at first glance."
Technology providers with superior onboarding and implementation processes, and with the resources to support implementation against these criteria, should be a requirement rather than a nice to have.
Lawyers have embraced AI, but hold valid reservations around the accuracy of its outputs. To scale up operations, organisational leaders will need to be transparent about how AI tools work, the source material driving them, and the risks that can crop up along the way.
Part 2
How lawyers are using AI
Most lawyers who use AI do so on a regular basis. Three-quarters (74%) told us they use it at least once a week, with 34% using it on a daily basis. Only 6% say they do not use AI tools at all.
Adoption is particularly high across in-house legal teams, with 89% using AI at least once a week, followed by those in mid-sized law firms (80%) and large law firms (78%). Regular use is lower among public sector respondents and small law firms (both 68%), academics (61%) and the Bar (52%).
Lawyers specialising in commercial & corporate and regulatory & compliance work are also more likely to be regular users, at 85%.
AI is being used most heavily for core legal tasks
Legal research is the most common use case for lawyers at 69%, followed by document summarisation at 62%, and drafting and document review at 53% each.
There were some noticeable variations in the data when looking at drafting alone.
Respondents from large law firms and in-house legal teams were more likely to use AI for drafting (66%). Leaders were also more likely than juniors to use AI for drafting (62% vs 51%) and legal research (73% vs 68%). Respondents working in regulatory & compliance (64%), employment (63%) and commercial & corporate (60%) were also more likely to use AI for drafting.
Lee from Simmons & Simmons believes organisations should prioritise the AI use case over the specific tool, and implement regular monitoring of how AI systems are performing over time.
"The most mature organisations are moving away from governing individual AI tools and towards governing the business use cases, processes and decisions that AI supports, ensuring accountability remains with humans throughout the lifecycle."
Patel from Eversheds advocates prioritisation when reviewing AI outputs based on the level of risk involved.
"High-risk use cases, particularly those involving client data, confidential information, legal advice or autonomous functionality, need more scrutiny."
Low-risk uses may be capable of moving more quickly, provided the right guardrails are in place, she says.
Again, the effectiveness of these use cases comes down to the quality of source material.
An in-house legal leader remarked that AI enables their team to draft work quicker but the human review process can be time-consuming.
"Sometimes reviewing AI outputs can take as long as if you drafted it yourself.”
New habits are forming around AI
One in 10 (10%) lawyers told us they are now dependent on AI to do their jobs. While still only a fraction, this shows a clear shift towards AI-dependence.
Dependence is highest in large law firms (15%) and in-house legal teams (13%), compared with 9% in small law, 7% in the public sector and 6% at the Bar.
Dependence is also higher among commercial & corporate (14%) and regulatory & compliance lawyers (12%), but lower in property (5%), family (6%) and personal injury & clinical negligence (4%).
Used well, AI can support junior development as an on-demand tutor, says Beckhaus from Freshfields.
"AI can explain the purpose and background of a clause, identify the interests of the parties, show how different provisions interact and help junior lawyers understand the wider commercial context."
"This can accelerate learning and reduce the repetition needed to recognise patterns and develop a deeper understanding of the subject matter."
Lawyers aren’t simply using AI to write emails or proofread documents. They’re increasingly relying on it for substantive legal work, from drafting contracts and answering stakeholder questions to identifying relevant cases and precedents. That reliance is growing, but it will only continue if trust grows with it.
Part 3
Scaling up on AI
Another barrier to scaling AI sits within organisations themselves.
Less than a third (29%) of respondents told us AI is already embedded in their organisation's strategy and operations, with the largest group (43%) saying they are experimenting but progress is slow.
A further 14% report interest but little investment, while relatively few describe outright resistance (6%) or say AI is not discussed at all (3%).
Overall, the picture is less one of resistance to AI and more of organisations struggling to move from experimentation to full-scale implementation. Unsurprisingly, large law firms lead the way on AI maturity, with 57% saying AI is embedded in strategy and operations, followed by in-house at 43%. This falls to only 13% for small law firms and just 3% for the Bar.
The majority of lawyers now believe AI will impact law firm billing models
The billable hour is unlikely to disappear any time soon, yet a growing number of respondents expect AI will have some impact on how law firms bill. More than half (55%) believe AI will change how firms bill, up from 40% in January 2025, with the figure rising consistently across each survey since then.
This is particularly true for large law firms and in-house legal teams, both at 62%.
Eversheds' Patel says:
"Clients expect firms to innovate and to use technology where they can to improve efficiency, quality and value."
By practice area, expectations are highest in regulatory & compliance (63%) and commercial & corporate (60%), but lower in property (44%), private client (46%) and family (33%).
Leaders are also more likely than juniors to expect billing change (60% vs 52%), suggesting they may be more attuned to AI’s commercial impact.
The biggest challenge is keeping pace with new technology
When asked about the biggest challenge their team is up against, the top response was keeping pace with new technology (53%). This was especially true for large law firm respondents (71%) and in-house teams (60%).
Other common responses were increasing costs (48%) and compliance demands (44%).
Lawyers are becoming more aware of the impact AI could have on their future careers.
Another consequence of a slow AI strategy is its impact on talent retention.
Nearly half (46%) of lawyers say their career would be negatively affected if their organisation failed to fully embrace AI, up from 28% this time last year.
Meanwhile, 22% say they would consider leaving an organisation that failed to embrace AI, up from 16% a year ago.
The gap suggests that lawyers increasingly see AI capability as important to their professional development.
Pressure is strongest in large law firms and in-house teams, where around six in ten say weak AI implementation could hurt their career, while around a third say they would also consider leaving.
In today's job market, lawyers who can demonstrate a strong understanding of AI in their workflows have a competitive advantage, one associate at a large law firm told us.
"Using AI for legal work shows a natural curiosity and a willingness to embrace change."
Scaling AI is no longer just about adoption. It is reshaping organisational culture, commercial models, day-to-day challenges and expectations around career development.
Firms are under pressure to move beyond experimentation, rethink how AI may affect billing, keep pace with rapidly changing technology and build an environment where lawyers feel they are developing the skills they will need for the future. Those that fail to do so risk falling behind both commercially and culturally.
AI adoption is accelerating, but trust will determine how far it can scale.
Yes, lawyers are already using AI for core legal work, but the impact will span to workflows, billing models, culture and skillsets.
The next phase will be less about whether legal teams adopt AI and more about how effectively they embed it. That means combining reliable technology and trusted legal sources with strong governance, human oversight and a culture that gives lawyers confidence to use AI well.
This report was edited by:
Dylan Brown
Content & Thought Leadership at LexisNexis UK
My role is dedicated to reporting on the big topics impacting law firms and legal teams. I run extensive surveys, speak with legal leaders and research market developments to produce reports you hopefully find helpful.