AI in Law Adoption Statistics 2026

AI in law statistics: how to read the 2026 adoption picture

AI adoption in legal work cannot be reduced to one universal percentage. Surveys measure different populations and ask different questions. Some count any personal use of an AI tool. Others ask whether a law office has adopted AI-based technology or whether an organization has incorporated generative AI into its work.

The strongest AI in law statistics therefore need their population, field period, and definition attached. The evidence below uses published primary research from the American Bar Association (ABA), Thomson Reuters Institute, and Clio. It separates office adoption from individual use and distinguishes general-purpose AI from legal-specific products.

The short answer is that experimentation is widespread, formal deployment is lower, and larger firms generally report more adoption. Governance, training, measurement, and client communication have not kept pace.

Headline legal AI adoption rates

The three major studies do not conflict as much as their headline numbers suggest. They measure different things.

The ABA 2024 Artificial Intelligence TechReport gives the narrowest and most transparent denominator. Of 512 attorneys answering the online-research volume, 30.2% said their offices currently used AI-based technology tools. The survey was administered from October through December 2024.

The Thomson Reuters 2025 executive summary reports organization-level generative AI adoption. Its early-2025 survey covered more than 1,700 professionals in the United States, United Kingdom, and Canada. It found use at 28% of law firms and 23% of corporate legal departments.

The Clio 2025 Legal Trends Report uses a broader individual-use measure. It found that 79% of legal professionals used AI in some capacity. This can include informal use of a general-purpose product even when the firm has not formally bought or governed an AI system.

Adoption by law firm size

The ABA data shows a clear size gradient. Large firms were much more likely than solo practices to report office use of AI tools.

These percentages come from the same 512-respondent ABA survey, so they are more suitable for a firm-size comparison than figures taken from separate studies. The difference between the largest firms and solo practices was 30.1 percentage points.

The result does not prove that firm size causes adoption. Larger firms can have more capital, technology staff, client pressure, and document volume, but the ABA survey is descriptive rather than causal.

General-purpose AI versus legal-specific tools

Clio's 2025 data helps explain why individual usage can be high while formal office adoption remains lower. It reported that:

• 79% of legal professionals used AI in some capacity.
• 46% used general-purpose AI tools such as ChatGPT, up from 36% in the prior year.
• 40% used legal-specific AI, down from 58% in the prior year.
• 82% expected their AI use to increase during the following 12 months.

These numbers describe categories that can overlap. A respondent can use both a general-purpose and a legal-specific tool, so the percentages should not be added.

The ABA asked a related but different question about products that firms had adopted or were seriously considering. ChatGPT led at 52.1%, followed by Thomson Reuters CoCounsel at 26.0% and Lexis+ AI at 24.3%. Those are consideration-or-adoption shares among the relevant respondents, not market shares and not proof of paid deployment.

What legal professionals use AI for

Thomson Reuters identifies document review as the most commonly cited current legal GenAI use case in its 2025 workflow analysis. Other prominent applications include legal research, document summarization, contract drafting and analysis, correspondence drafting, and knowledge retrieval.

The ABA data focuses on perceived benefits rather than verified output gains. Among its respondents:

The wording matters. A perceived benefit is not a measured return. It shows what attorneys expect from AI, not how many hours or dollars a deployed tool has saved.

For firms evaluating which work should remain under direct human control, our comparison of AI and human virtual assistants provides a broader operations view.

Accuracy, privacy, and reliability concerns

The ABA survey found that risk concerns remain common even as use grows.

These were concerns respondents could identify, not measured incident rates. The figures do not mean that 74.7% of tools produced an accuracy failure. They mean 74.7% of surveyed attorneys identified accuracy as a concern.

Legal organizations also have professional duties that make an ordinary consumer-tool rollout insufficient. Inputs can contain privileged or confidential information, outputs can invent citations, and a plausible summary can omit controlling authority. Human review, approved data handling, source checking, and clear responsibility remain necessary.

Policy and training lag behind usage

Thomson Reuters found a notable governance gap in its early-2025 sample:

• 52% said their organization did not have a GenAI usage policy.
• 64% had received no specific GenAI training.
• Only 20% of surveyed legal professionals measured return on investment.
• 71% of law firm clients did not know whether their firms used GenAI.

The first two figures cover the broader professional sample described by Thomson Reuters, not only attorneys. The ROI figure covers legal professionals in the report. The client-awareness figure comes from surveyed law firm clients. Keeping those populations distinct prevents a broad survey result from being presented as a universal law-firm rate.

This gap is operational as well as ethical. A firm cannot consistently control confidential inputs, validate outputs, or explain value to clients if tool use remains informal and staff receive no role-specific guidance.

Expected adoption over the next five years

Expectations are strong, but they are forecasts rather than observed adoption. Thomson Reuters reported that 95% of professionals in its survey expected GenAI to become central to their organization's workflow within five years.

The ABA survey asked attorneys when AI would become mainstream in legal work:

The results show strong expectations, not a guaranteed timeline. Procurement, accuracy, client rules, regulation, and integration with legal systems can all affect actual adoption.

A practical adoption benchmark for law firms

The published research supports a four-part benchmark:

1. Usage: Track the share of lawyers and staff who use approved AI tools, separated by general-purpose and legal-specific systems.
2. Workflow: Record which tasks use AI and which require a lawyer's direct review or approval.
3. Governance: Measure policy coverage, training completion, data controls, and source-validation compliance.
4. Outcomes: Measure cycle time, rework, client satisfaction, cost, and quality before and after deployment.

This approach avoids treating a login as successful adoption. It also makes comparisons across teams more useful. A firm can have high individual use but weak governance, or a licensed legal AI platform with little repeat workflow use.

Administrative workflows are a practical place to distinguish legal judgment from repeatable support. See legal administrative assistant services and virtual legal assistant services for examples of human-supported legal operations.

What the AI in law statistics show for 2026

The evidence available in 2026 shows a profession between experimentation and formal adoption. In late 2024, 30.2% of ABA respondents said their offices used AI-based tools. In early 2025, Thomson Reuters measured GenAI use at 28% of law firms and 23% of corporate legal departments. Clio's broader individual-use measure reached 79% in 2025.

Firm size, tool type, and survey wording explain much of that range. The safest conclusion is not that one number is correct and the others are wrong. It is that personal experimentation has spread faster than organization-wide deployment, legal-specific integration, policy, training, and ROI measurement.

Frequently asked questions

What percentage of law firms use AI?

The answer depends on the population and definition. Thomson Reuters reported that 28% of law firms used generative AI in early 2025. The ABA found that 30.2% of 512 surveyed attorneys said their offices used AI-based technology tools in October to December 2024.

Do larger law firms adopt AI faster?

In the ABA's 2024 sample, yes. Reported office use ranged from 17.7% among solo attorneys to 47.8% at firms with 500 or more lawyers. The survey shows an association with firm size, not proof that size causes adoption.

What is the biggest concern about legal AI?

Accuracy was the most selected concern in the ABA survey at 74.7%. Reliability followed at 56.3%, and data privacy and security at 47.2%. These are reported concerns, not incident rates.

Is AI use in law formally governed?

Often not. In Thomson Reuters' early-2025 professional-services survey, 52% said their organization lacked a GenAI use policy and 64% had not received specific GenAI training.

Sources and methodology

• American Bar Association, 2024 Artificial Intelligence TechReport, based on the Legal Technology Survey administered from October through December 2024; the online-research AI results cover 512 attorney respondents.
• American Bar Association, 2024 Legal Technology Survey overview, which describes the five-volume private-practice survey program.
• Thomson Reuters Institute, 2025 GenAI report executive summary for legal professionals, published April 15, 2025; more than 1,700 respondents in the US, UK, and Canada surveyed in early 2025.
• Thomson Reuters Institute, legal GenAI workflow analysis, published May 29, 2025.
• Clio, 2025 Legal Trends Report, with methods and population details in Appendix B.

Percentages are reproduced with the population and question context available in each source. They should not be combined into a single blended adoption rate. This article was last verified on September 4, 2026.