AI Use in Legal Work Has Nearly Doubled: What the Data Shows About GenAI Adoption

Artificial intelligence is moving rapidly from experimentation to regular legal work. GenAI use among surveyed legal professionals increased from 14% in 2024 to 26% in 2025. The rise shows that AI is increasingly being used for document review, legal research, summarisation, drafting and other routine legal tasks, while questions around accuracy, confidentiality, training and professional responsibility remain important.

AI Use in Legal Work: 14% in 2024 to 26% in 2025

The most striking figure from the 2025 Thomson Reuters Generative AI in Professional Services report is the speed at which actual GenAI use has increased.

Add LawBhoomi as your preferred source on Google.

Add Now →

Among surveyed legal professionals:

• 14% reported using GenAI in 2024
• 26% reported using GenAI in 2025

This represents an increase of 12 percentage points in a single year. In relative terms, the number of professionals reporting GenAI use increased by approximately 86%, meaning adoption came close to doubling within a year.

The distinction between percentage-point growth and percentage growth is important. Adoption did not increase by merely 12%. It moved from a base of 14% to 26%, which represents a much larger proportional increase.

The figures indicate that GenAI is increasingly becoming part of actual legal workflows rather than remaining a subject of experimentation or discussion.

How Widely Is GenAI Being Used in the Legal Sector?

The 2025 data suggests that legal professionals are among the stronger adopters of generative AI within professional services.

GenAI adoption was reported at:

• 28% among law firm professionals
• 23% among professionals in corporate legal departments

This indicates that law firms were slightly ahead of in-house legal teams in reported GenAI adoption.

The gap is not particularly large, however. Both figures show that artificial intelligence has entered mainstream legal environments, including private legal practice and corporate legal departments.

The shift is particularly significant because legal work traditionally depends heavily on professional judgement, confidentiality, precedent, interpretation and detailed review. Adoption of AI in such an environment suggests that legal professionals are increasingly identifying specific tasks where technology can assist without necessarily replacing professional decision-making.

Is AI Becoming Part of Everyday Legal Practice?

Adoption statistics alone do not show how deeply technology has entered professional routines.

Frequency-of-use data provides a clearer answer.

Among existing GenAI users surveyed:

• 72% used GenAI at least once a week
• more than 40% used it once a day or multiple times a day

Relatively few users reported using the technology only once or twice a month.

This suggests that once legal professionals begin integrating GenAI into suitable workflows, its use can become regular rather than occasional.

The development is important because technological transformation often happens in stages.

A professional may initially experiment with AI for summarisation. Once confidence develops, the same technology may be used for legal research, drafting, document comparison and internal knowledge management.

The shift from experimentation to repeated use is therefore one of the strongest indicators that GenAI is becoming operational technology within legal practice.

What Are Lawyers Using Generative AI For?

The data becomes more significant when the actual use cases are examined.

Among legal professionals using GenAI, the six leading applications identified in the report were:

Document Review — 74%

Document review was the most commonly reported GenAI use case, at 74%.

Legal practice frequently involves analysing large volumes of documents. Litigation may require the examination of pleadings, correspondence, contracts, reports and evidentiary materials. Transactional work can similarly involve extensive due diligence.

AI tools can assist in identifying relevant clauses, extracting information, organising documents and highlighting potentially important material.

This does not eliminate the need for legal review. Instead, it can reduce the amount of time spent on the initial examination of large document sets.

Legal Research — 73%

Legal research was reported as a GenAI use case by 73% of legal professionals using the technology.

Legal research has traditionally involved searching cases, statutes, regulations, commentaries and secondary materials.

Generative AI can make the preliminary research process faster by helping identify issues, generate possible research directions, summarise legal materials and organise information.

However, legal research remains an area where accuracy is particularly important. AI-generated citations, propositions and summaries cannot automatically be treated as authoritative legal sources.

Primary authorities such as judgments and legislation continue to require verification.

Document Summarisation — 72%

Document summarisation was used by 72% of respondents using GenAI.

This is one of the most natural applications of generative AI in legal practice.

Long judgments, agreements, regulatory documents and case files often contain hundreds of pages. AI systems can assist in producing preliminary summaries, identifying major issues and extracting important provisions.

For lawyers handling large volumes of information, summarisation can significantly reduce the time required to understand the broad structure of a document.

The final interpretation, however, remains a professional legal function.

Brief or Memo Drafting — 59%

About 59% of legal professionals using GenAI reported using it for brief or memorandum drafting.

AI can assist with:

• preparing an initial structure for a legal memorandum;
• organising arguments according to identified legal issues;
• converting research into a preliminary draft;
• rewriting complex material in clearer language; and
• summarising authorities for further review.

Drafting is nevertheless more sensitive than simple summarisation because legal documents may directly affect client rights, litigation strategy and professional liability.

Human review therefore remains essential.

Contract Drafting — 51%

Contract drafting accounted for 51% of reported legal GenAI use cases.

AI tools can help create first drafts, suggest clauses, compare provisions and identify inconsistencies.

The technology may be particularly useful for repetitive contractual work involving standard structures.

However, contracts cannot be treated simply as collections of standard clauses. Commercial context, bargaining power, applicable law, regulatory obligations and risk allocation can materially change how a provision should be drafted.

AI-generated contract language therefore requires careful legal analysis.

Correspondence Drafting — 50%

Half of legal GenAI users surveyed reported using the technology for correspondence drafting.

This may include preliminary drafts of client communications, internal memoranda, routine notices or professional correspondence.

Such use demonstrates an important pattern in AI adoption: professionals may initially find the greatest value in tasks that are frequent, repetitive and time-consuming rather than in functions involving final legal judgement.

Why Is AI Adoption Increasing in the Legal Profession?

The principal reason identified by professionals is straightforward: time and productivity.

Legal practice contains numerous tasks requiring significant amounts of reading, comparison, drafting and information processing.

Generative AI is particularly suited to assisting with these activities.

Faster Processing of Large Amounts of Information

Lawyers regularly deal with information-heavy assignments.

A single matter may involve:

• hundreds of pages of agreements;
• large sets of correspondence;
• multiple judgments;
• regulatory materials;
• evidence and annexures; and
• previous versions of legal documents.

AI can assist with the first layer of processing, allowing relevant material to be identified faster.

Reduced Time on Repetitive Work

Certain legal tasks require professional attention but involve considerable repetition.

Examples include extracting contractual provisions, preparing first drafts, organising chronologies or summarising similar documents.

Automation can reduce the amount of time spent on these processes.

More Time for Higher-Value Legal Work

The strongest argument for legal AI is not necessarily that machines will replace lawyers.

Rather, technology may change how lawyers allocate professional time.

If routine information processing can be completed faster, additional time may become available for:

• legal strategy;
• client counselling;
• negotiation;
• advocacy;
• factual analysis;
• risk assessment; and
• complex legal reasoning.

These functions depend heavily on judgement and contextual understanding.

Do Legal Professionals Believe AI Should Be Used?

Growing use is accompanied by increasingly positive attitudes towards GenAI.

The findings show that 59% of surveyed law firm professionals and 57% of corporate legal professionals believed GenAI should be applied to their work.

Across surveyed professional sectors, 55% of respondents said they were hopeful or excited about the future of GenAI. This represented an increase of 11 percentage points from 2024.

These figures suggest that the debate is gradually shifting.

The question is becoming less about whether artificial intelligence will have any role in legal practice and more about the conditions under which it should be used.

What Are the Main Risks of Using AI in Legal Work?

Rapid adoption does not mean that the risks surrounding artificial intelligence have disappeared.

For legal professionals, several concerns are especially important.

Accuracy and Hallucinated Information

Generative AI can produce information that sounds authoritative even when it is inaccurate.

In legal work, such errors can be serious.

A fabricated judgment, incorrect statutory provision or inaccurate statement of law may affect legal advice or court proceedings.

AI-generated legal research therefore requires verification against reliable legal authorities.

Confidentiality

Legal professionals regularly handle sensitive information relating to clients, litigation, transactions and investigations.

Entering confidential information into inappropriate AI systems may create privacy, confidentiality or data-security risks.

Law firms therefore need clear rules governing which tools can be used and what information may be processed through them.

Professional Responsibility

Legal practitioners remain professionally responsible for work submitted in their name.

Using an AI-generated draft does not transfer responsibility for its accuracy to the software.

The lawyer remains responsible for checking:

• facts;
• legal authorities;
• citations;
• statutory provisions;
• contractual language; and
• conclusions.

Bias and Incomplete Context

AI-generated responses are influenced by their underlying data and system design.

Legal questions often depend on factual nuances, jurisdiction, procedural history and competing authorities.

A technically plausible answer may therefore still be legally incomplete.

Human legal reasoning remains necessary to determine whether an AI-generated output properly reflects the applicable law and facts.

Are Law Firms Ready for Widespread AI Adoption?

The adoption figures reveal an important contradiction.

Use of AI is increasing rapidly, but institutional preparation has not always increased at the same pace.

The report found that:

• 52% of surveyed professionals said their organisation did not have a GenAI usage policy
• 64% had received no specific GenAI training

At the same time, access to training was improving. Around 31% reported having access to training in 2025, compared with 19% in 2024. Training availability reached 40% among law firm respondents and 41% among corporate legal practitioners.

This creates one of the most important issues in legal AI adoption.

Technology can spread faster than organisational governance.

Without clear institutional policies, different professionals may use AI tools in inconsistent ways. Some may enter confidential information into unsuitable platforms, while others may rely on generated material without sufficient verification.

AI governance is therefore becoming part of legal risk management.

What Should an AI Policy for Legal Work Address?

A legal organisation introducing GenAI should ordinarily develop clear internal standards covering its use.

Important areas include:

Approved AI Tools

Organisations should identify which systems may be used for professional work and whether those systems satisfy appropriate privacy and security standards.

Confidential Information

Policies should explain whether client information, privileged communications or commercially sensitive materials can be entered into AI systems.

Verification Requirements

AI-generated legal propositions, authorities and citations should be independently checked before professional reliance.

Human Review

Documents generated or materially assisted by AI should remain subject to appropriate professional supervision.

Training

Legal professionals need to understand both the capabilities and the limitations of the tools being used.

The objective is not merely to teach software operation. Training must also address professional judgement, confidentiality, accuracy and responsible use.

Are Clients Expecting Law Firms to Use AI?

Another important development is emerging from the client side.

According to the report:

• 59% of corporate legal clients wanted their outside law firms to use GenAI
• 44% of government legal clients expressed the same preference

At the same time, 71% of law firm clients did not know whether their external firms were using GenAI.

This creates an interesting gap between client expectations and law-firm communication.

Clients may increasingly expect technology to improve efficiency. However, firms also need to explain how AI is being used, what safeguards apply and whether technological efficiency affects pricing or service delivery.

Is AI Changing the Law Firm Business Model?

The relationship between AI and legal fees is likely to become increasingly important.

Traditional law-firm economics have often been connected to billable hours. However, if technology reduces the time required to perform certain tasks, clients may question whether billing models should change.

The 2025 findings indicate that the industry has not reached a settled position.

Among surveyed legal firms, 53% believed rates would remain unchanged, while 20% expected rates to increase somewhat or significantly despite increasing AI adoption.

Another significant finding is that only 20% of surveyed legal professionals were measuring the return on investment from GenAI.

This suggests that many organisations are adopting AI before developing sophisticated systems for measuring its financial value.

Useful measures could include:

• time saved on particular tasks;
• cost savings;
• turnaround time;
• accuracy improvements;
• lawyer productivity;
• client satisfaction; and
• profitability of matters.

Without such metrics, the actual commercial impact of AI may remain difficult to determine.

Will AI Replace Lawyers?

Current data does not support a simple conclusion that AI is replacing lawyers.

The strongest current use cases involve assisting lawyers with information-heavy tasks, including research, document review, summarisation and drafting.

These functions form important parts of legal practice, but legal services involve considerably more than producing text.

Professional legal work may require:

• understanding a client’s commercial or personal objectives;
• determining which facts are legally relevant;
• assessing conflicting evidence;
• interpreting uncertain legal rules;
• exercising professional judgement;
• developing litigation strategy;
• negotiating settlements;
• advising on risk; and
• representing clients before courts and authorities.

AI may affect how these services are delivered, but responsibility for professional judgement continues to remain with legal professionals.

A more realistic development is the transformation of legal workflows rather than the complete replacement of lawyers.

What Skills Will Lawyers Need in an AI-Driven Legal Profession?

The increasing adoption of GenAI may gradually change expectations concerning legal skills.

Traditional legal abilities will remain important, particularly legal research, drafting, interpretation, advocacy and reasoning.

However, additional abilities are likely to become valuable.

AI-Assisted Legal Research

Legal professionals will need to understand how to obtain useful preliminary research from AI while independently verifying authorities.

Better Review Skills

If machines produce first drafts faster, the ability to critically review those drafts becomes more important.

Prompting and Instruction

AI systems respond according to the instructions and context provided.

Professionals who understand how to structure legal questions clearly may obtain more useful outputs.

Technology Risk Awareness

Lawyers increasingly need to understand issues relating to privacy, confidentiality, data security and AI-generated errors.

Professional Judgement

Perhaps paradoxically, greater automation may increase the importance of human judgement.

When routine drafting becomes easier, the ability to determine whether a draft is strategically, commercially and legally appropriate becomes more valuable.

The findings already show that 21% of law firm respondents regarded GenAI skills as a “nice-to-have” qualification when evaluating potential recruits.

What Does the Future of AI in Legal Work Look Like?

The expectation of further adoption is exceptionally high.

The report found that 95% of surveyed professionals expected GenAI to become a central part of their organisation’s workflow within the next five years.

This does not mean every legal task will become automated.

It does indicate that AI-assisted workflows are likely to become increasingly normal across areas such as research, litigation support, contract management, knowledge management and drafting.

The next stage may also involve greater use of agentic AI, where AI systems perform several connected tasks with comparatively less human intervention before presenting work for professional review. The report identified this as an emerging development in the legal AI landscape.

Such systems may further increase the importance of governance and supervision.

What Does the Rise From 14% to 26% Really Mean?

The movement from 14% GenAI adoption in 2024 to 26% in 2025 is more than a technology statistic.

It reflects a broader change in how legal services are being performed.

AI is already assisting professionals with some of the most time-consuming elements of legal work. Document review, research and summarisation all have adoption rates exceeding 70% among surveyed legal GenAI users. Drafting is also becoming a major area of application.

At the same time, the data exposes a preparedness gap. Many organisations still lack formal AI policies, many professionals have received no specialised training, and relatively few organisations measure AI’s return on investment.

The next phase of legal AI adoption is therefore likely to focus less on whether the technology can be used and more on how it can be used safely, accurately and efficiently.

Conclusion

Artificial intelligence is becoming an increasingly important part of legal work. The rise in reported GenAI use from 14% in 2024 to 26% in 2025 shows that adoption is moving rapidly, particularly in document review, legal research, summarisation and drafting.

The data does not suggest that professional legal judgement has become less important. Instead, it indicates that routine and information-intensive components of legal work are increasingly being assisted by technology. The long-term impact will depend on how effectively law firms and legal departments combine AI efficiency with human supervision, confidentiality, accuracy, ethical responsibility and professional expertise.