One policy cannot anticipate every way lawyers will use AI. That is why governance needs a feedback loop. What lawyers learn from everyday use should shape the guidance they follow.

The SRA has warned about false legal material and risks to client confidentiality. As AI becomes further ingrained in everyday legal work, those risks sit alongside questions of trust and accountability.

Lawyers need to know which tools to use, what to check before relying on an answer and where to take a concern. When lawyers spot a problem, someone needs to assess it, change the guidance and tell the wider team.

That feedback loop is what keeps governance useful as the tools change.

Our latest survey found:

Three-quarters of UK lawyers use AI at least once a week

Concerns about inaccurate or fabricated AI information are growing

AI grounded in legal sources makes lawyers more confident when using it

Despite high adoption, most organisations have yet to embed AI in their strategy and operations

AI ownership is often split across multiple teams and functions

In this report

We examine how AI governance is taking shape today, where the main gaps are emerging and how governance may need to evolve as AI capabilities grow.

1. Lawyers are using AI for core legal activities

Governance needs to look at each AI use case in context

Our July 2026 survey of 543 UK lawyers and legal professionals found three-quarters use AI at least once a week, and more than a third use it every day.

Legal research is the leading use case among UK legal professionals (69%), followed by document summarisation (62%). Drafting and document review were each used by more than half of respondents (53%). These patterns are similar across the market. Weekly AI use reaches 89% among in-house lawyers, 80% in mid-sized firms and 78% in large firms.

While AI usage is high, it is moving faster than organisational governance. Just 29% of UK legal professionals say AI is embedded in strategy and operations while almost one in five cannot identify a clear owner for AI governance.

Lawyers may already be using tools before clear rules are in place on suitable tasks, data use, review and accountability. Dan Murphy, Principal at DJ Murphy Solicitors in Cardiff, warns,

"The long term risks of an ineffective legal AI governance strategy include the loss of trust, reputation and degradation of skills."

Katie Griffis, General Counsel at Control Risks, adds:

"If clients trust the organisation and they use AI irresponsibly, their reputations may never recover. They might actually be more at risk."

That gap can leave the benefits with the individual while the wider risks sit with the organisation.

The governance question now is how these tools are being used, and what's at stake when their output shapes both internal and client-facing work. The level of oversight should reflect the sensitivity of the work, the role AI plays and the consequence of an error. For instance, low-risk internal work might need lighter review, yet work involving client data, confidential information, legal advice or greater autonomy calls for closer scrutiny. As Peter Lee, Partner and Global Head of AI Governance at Simmons & Simmons says,

"Organisations should start by asking what problem they are trying to solve and whether AI is genuinely the right solution."

The harder questions emerge when looking at output. Who decides whether a summary is good enough? Must every cited authority be opened and checked? What happens if a tool becomes unavailable midway through a matter? Can the team reconstruct how an answer was reached?

Governance frameworks must answer those questions in a form people can apply while the work is being done. A practical risk lens means matching the level of governance to the level of risk. Lower-risk uses may only need basic checks, while higher-risk work, such as client advice or confidential matters, needs stronger controls, documented review and clear escalation.

A risk-based structure gives people clearer boundaries for using AI in their day-to-day work. It allows teams to move more quickly where the consequences are limited and apply closer scrutiny as the stakes increase. As Lee explains,

"People are generally more willing to embrace AI when they understand its purpose, its limitations and the safeguards surrounding its use."

Those safeguards depend in part on the tools themselves. Teams need to understand what a tool can access, how it handles data and whether its answers can be checked before deciding where it belongs in legal work.

2. Your governance comes down to your tools

The foundations of your AI governance strategy are built on the tools your team uses

The focus has switched from if you use AI to how you use AI.

The challenge many organisations now have is whether the tools being used can meet their standards.

That assessment requires a clear understanding of how the technology works, uses data and produces outputs. Bhavisa Patel, Director of Technology Products at Eversheds Sutherland, explains:

"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."

Concern about AI accuracy has risen alongside adoption. Our 2026 survey revealed that 83% of respondents were concerned about inaccurate or fabricated AI information, up from 57% when LexisNexis UK began measuring the issue in early 2024. As AI is used across more substantive legal tasks, the ways in which errors can enter legal work become easier to see.

Provider selection is one place where those concerns become practical. When respondents were asked what they look for in a legal AI provider, data privacy and security standards ranked first among both private practice and in-house respondents, cited by 55% and 64% respectively. Transparency about how providers use AI also ranked highly, at 38% in private practice and 41% in-house. Human oversight of AI-enabled services was cited by 40% of private practice respondents, while 40% of in-house respondents selected cost and commercial value.

The work a tool will support shapes how it should be assessed. A firm using AI to help prepare client advice and an in-house team using it to advise a business may need different review processes and measures of value. For example, Kate Finch, General Counsel at Access Fertility, says:

"An ineffective legal governance strategy can create compounding risks across compliance, client trust and operational resilience. Over time, this may lead to regulatory scrutiny, inconsistent decision-making, data misuse and reputational damage that is far harder to repair than the original governance gap. Organisations that fail to govern AI and legal processes well may also lose commercial ground to competitors that can offer more reliable and defensible services."

The question for each team is whether the provider’s data handling, outputs and safeguards meet the demands of its particular use case.

The sources behind an AI tool are part of that assessment, with 81% of respondents saying they feel more comfortable using AI when it's grounded in legal sources. Tony Randle, Director of Client Tech and Service Improvement at Shoosmiths, expands on the importance of source reliability:

"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."

For organisations, that extends governance beyond deciding which products are approved. It requires scrutiny of the information those tools draw on, what users can verify, how client and organisational data is handled, and what controls support their use.

Who should be responsible for AI governance?

Internal governance determines how AI enters legal work. In many organisations, responsibility is already spread across several functions. IT and technology teams are most commonly involved (38%), followed by risk and compliance (34%) and executive leadership (31%). In-house legal teams also play a substantial role within corporate legal functions (24%).

The survey found nearly four in ten (39%) legal professionals identified multiple teams or functions as responsible for AI governance, while 40% identified only one clear owner. A further 18% said there was no owner at all.

That breadth reflects the range of expertise AI governance can require. As Patel explains:

“Strong relationships between technology, compliance, legal, risk and information security teams are critical, because open communication makes it much easier to balance business value with responsible adoption.”

Lawyers need to be able to check where an AI answer comes from and compare it with trusted legal sources. When something goes wrong, that should help improve how AI is used across the organisation.

Randle describes effective AI governance as this:

"It is a live feedback loop between lawyers at the coal face, risk & compliance teams, technology departments and regulators.”

Yet governance is developing unevenly across the profession. Almost one in five UK legal professionals say there is no clear owner for AI governance in their organisation, with the figure rising among those working in small firms, at the Bar and in academic institutions.

From shared responsibility to clear ownership

Where governance spans several functions, clear coordination helps turn shared expertise into clear decisions. Named leaders can give AI governance conversations a consistent centre. Lee describes the elements that support ongoing governance:

“In practice, organisations need clear ownership, proportionate risk assessment, oversight of high-impact use cases, and regular monitoring of how AI systems are performing over time.”

Leadership also shapes how those responsibilities are understood and applied. As Patel explains:

"Leaders will need to embed a mindset where people understand the strengths and limitations of AI, feel confident using it responsibly, and know how to review and refine outputs."

In practice, that collaboration should leave legal professionals with clear answers about what they can use, what they can enter, how outputs should be checked, who owns the final work and where concerns should go. Five practical questions can give that shared governance a common reference point:

Consistent answers turn collective input into clear day-to-day decisions and give human oversight practical meaning: who checks, what they examine, when they intervene and how concerns escalate.

The next challenge is keeping that structure useful as tools, use cases and expectations continue to evolve.

3. Keeping pace with technology is your biggest challenge

Governance must be embedded into operations and processes from day one

As AI tools develop, the most effective approach is to embed governance into existing processes rather than creating separate approval structures.

When evaluating the biggest challenges facing organisations, keeping pace with technology now sits ahead of familiar pressures such as rising costs, regulatory demands, changes in the law and cybersecurity. Yet just 29% of UK legal professionals say AI is embedded in their organisation’s strategy and operations, even as technology choices shape how legal work is delivered, how teams operate and where organisations invest.

Identifying AI initiatives early, directing them through appropriate assurance channels and embedding governance into existing processes will help drive innovation forward.

Commercial expectations are shifting at the same time. More than half of respondents, 55%, expected AI to change law firm billing, up from 40% in January 2025. As AI changes the time and process involved in producing legal work, firms are likely to face closer questions about value, quality and accountability.

As Patel says:

"Clients expect firms to innovate and to use technology where they can to improve efficiency, quality and value."

For individuals, the consequences are increasingly personal. Almost half of respondents believed their career would be harmed if their organisation failed to embrace AI fully, while close to a quarter would consider leaving. The figures are higher in some parts of the profession, including large law firms and in-house teams.

Together, these pressures are reshaping expectations across the profession. Clients are placing greater emphasis on efficiency, quality and value in how legal services are delivered; organisations are under pressure to keep pace with technology; and legal professionals increasingly see AI capability as relevant to their own careers.

Governance gives each of those groups a clearer basis for what responsible AI use should look like as the technology develops.

Build the governance model as AI capabilities develop

As AI tools support more of the legal workflow, organisations will face more decisions about where and how they should be used. A governance model established now gives leaders and teams a consistent basis for assessing new uses, setting the right level of review and deciding where professional judgement needs to remain central.

The aim is to put that decision-making structure in place so it can support the organisation as the range of possible uses develops. As Randle says:

"Governance needs to keep pace with the speed at which technology and its uses are evolving. That is a challenge, but one in which every firm must invest. That means governance must be a continuous process of testing, learning and updating, rather than just an annual exercise."

What should legal teams expect from AI providers?

AI vendors have a responsibility as well to design systems to make responsible use of AI the easy path. Providers should offer clear data safeguards, answers linked to sources lawyers can check, controls over who can use the tool and for what, and records that help teams review how it was used.

They should explain the tool’s limitations, respond to reported errors and tell customers when a change affects how the tool works. That gives legal teams something practical to build their own governance around.

When legal work shapes governance

AI is already relied upon for research, summaries, drafting and review. Almost three-quarters of UK legal professionals use it weekly, yet very few say it's embedded in their organisation’s strategy and operations.

AI governance frameworks need to cover how tools are chosen and how their use is managed.

Choosing a tool means understanding what information it can access, how it handles data and whether users can check the sources behind its answers. The safeguards should then reflect the work the tool will support, with closer scrutiny where confidential material or legal advice is involved.

Governance also needs a place in the processes teams already follow. Clear owners can decide which uses are approved, what review is required and how concerns are investigated. What lawyers discover in practice should feed into those decisions, changing guidance or tool choices when needed.

That is how governance can keep pace with the work it oversees.

Methodology

LexisNexis surveyed 543 legal professionals in July 2026. Trend comparisons use LexisNexis surveys from early 2024, July 2025 and January 2026. The 543 respondents included 43% from private practice, 27% in-house, 13% from the Bar, 13% from the public sector and 3% from academia. By seniority, 59% were lawyers from trainee to senior level, 34% were heads of department, partners, General Counsel or other senior leaders, and 6% were non-lawyers.

Practice note

The role of explainability in AI tools

Explainability has emerged as an important component of ethical and responsible artificial intelligence (AI) and a common requirement of developing AI legislation and regulation.

Produced in partnership with Minesh Tanna of Simmons & Simmons, William Dunning of Simmons & Simmons and Jacob Turner of Fountain Court Chambers.

Read the practice note here

Practice note

AI in the workplace

Artificial intelligence (AI) is being used with increasing frequency in employment and human resources (HR) contexts. Activities into which AI is being integrated occur throughout the employment lifecycle, covering recruitment, monitoring of employees and performance management.

Produced in partnership with Jonathan Newman of Simmons & Simmons and Olivia Galliano of Simmons & Simmons

Read the practice note here

Researched and written by:

Maya Sternthal

Brand Marketing Specialist at LexisNexis UK

I research and write about the major issues shaping the legal sector, with a focus on legal innovation, emerging technologies and the rule of law. My role involves sharing data-driven findings to market through reports, industry insights and content for legal professionals.

Read my articles here

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.

Read my articles here