by Nigel Balmer, Legal Services Research Centre

Few topics are generating as much debate as artificial intelligence. Will it transform legal practice or undermine it? Is it a tool for access to justice or a risky shortcut? More broadly, will we all be living on Caribbean islands with robot butlers, and if so, will we have enough ice to make frozen daiquiris? Views can be strongly held, but to date many have been shaped by anecdote, speculation and modest studies.

Our new report, Generative AI Use in the Legal Profession: Findings from the 2025 Victorian Lawyer Census, provides a large-scale quantitative picture of AI adoption across the Victorian legal profession. Drawing on responses from 1,887 lawyers, it examines who is using AI, what for, and what it means for the profession and those who depend on legal services.

Who is using AI, and how?

Just over a third (37%) reported using AI in their legal practice. Use varied across the profession; higher among younger lawyers and those speaking a language other than English at home, for example, and markedly lower among barristers (15%) and government practitioners (11%). General-purpose tools such as ChatGPT, Claude and Gemini, were more commonly used than purpose-built legal AI, with use primarily applied to background research, legal research and client communication rather than legal decision-making or court documents.
 

Barriers, quality assurance and guidance

Almost half of non-users cited data, privacy and security concerns as reason for not using AI, followed by the absence of clear guidelines. Among AI users, the vast majority (89%) reported taking steps to ensure accuracy of outputs.

Sixty per cent had not read existing court or regulator guidance on AI use, including close to half of those already using it, and workplace AI guidelines existed for only 42%. Half had received AI training in the past year, with rates far higher among users (72%) than non-users (38%).

Attitudes, risks and benefits

Nearly all (96%) agreed that lawyers have a duty to ensure AI use complies with professional obligations, and most indicated that use should be disclosed to clients and in litigation. Just over half considered AI a necessary element of modern practice, despite only a third currently using it.

Accuracy, privacy and security were the most widely recognised risks, while fears about AI replacing lawyers ranked lowest (33%). Non-users exhibited heightened risk perceptions on almost all dimensions. 

Perceived benefits were narrowly conceived, centring on enhanced efficiency (71%), with far fewer perceiving improved affordability (46%), accessibility (45%), quality (31%) or client satisfaction (26%). Exposure to training and workplace guidelines was associated with greater recognition of benefits, though without a corresponding increase in perceived risk.

Looking ahead

The report identifies three challenges: 

• AI use is concentrated among newer entrants to the profession, raising questions about its impact on the development of core professional skills, particularly where use extends beyond information-gathering to legal analysis.
• Lawyers in non-traditional settings may face tensions where enterprise AI guidelines conflict with professional obligations. This may be compounded by low awareness of existing guidance. 
• AI's benefits remain narrowly focused on efficiency for providers, with applications that could enhance accessibility, quality or affordability largely unexplored.

Realising AI's potential to make legal services more effective and equitable will require deliberate effort to broaden how its benefits are conceived and to whom they flow.

Read the full report, Generative AI Use in the Legal Profession: Findings from the 2025 Victorian Lawyer Census, and accompanying research brief. 

Sign up to our latest news and research for more like this and AI findings from the 2026 census as soon as they come out