Oct 30 edition/Reporting & analysis
BusinessPolicySafety

BusinessMarkets, money & strategy

Victorian lawyers' AI use rises to 55% in a year, but only a quarter of users say client fees fell

A Victorian Legal Services Board and Commissioner census of 2,221 lawyers found AI use rose from 37% to 55% in a year. Only 26% of users report lower client fees, and clients' own chatbot use often adds work.

From the The Guardian · AI article of October 29, 2026 and cited sources: More lawyers are using AI, and most use general chatbots. [2] [6]

SOURCES[2] VLSB+C: AI use by Victorian lawyers surges, but benefits… lsbc.vic.gov.au[6] VLSB+C: New research examines AI use in the Victorian… lsbc.vic.gov.au

AI-assisted, not human-edited. Methodology Corrections Who owns this site

KEY FACTS
What happened
More lawyers are using AI, and most use general chatbots. [2] [6]
Original source
The Guardian · AI article, independent perspective
Brief by
AT Pulse research desk AI-assisted; publisher and contact on the desk page
When
October 29, 2026 article published
Evidence
7 sources across 4 sites

Reach: not measured yet

Illustration from the Guardian: Victorian lawyers' AI use rises to 55% in a year, but only a quarter of users say client fees fell
Image: the Guardian — Original article ↗
THE CORE IDEAS4 TAKEAWAYS
01

More lawyers are using AI, and most use general chatbots. Use rose from about 37% in 2025 to 55% in 2026, and 78% of users say they now use it more. About 79% use general-purpose tools such as ChatGPT, Claude, Gemini or Copilot. [2] [6]

02

Clients see few of the savings. Only 26% of AI users say fees fell, with 34% reporting cuts for business clients and 23% for individuals. Very few changed how they charge: 7% moved to fixed fees and 3% unbundled their services. This suggests most efficiency gains stay inside hourly billing. [1] [7]

03

Clients' own AI use often creates more work for lawyers. Among client-facing lawyers, 46% say it adds work and 19% say it reduces work. 41% have received AI-generated documents that needed correcting, and only 10% say clients understand their legal problem better because of AI. Community legal centres report being hit especially hard. [2] [7]

04

Courts and the regulator are tightening controls. A Victorian lawyer was sanctioned for filing false AI-generated citations. The Federal Court issued a practice note that can require disclosure of AI use. The High Court's chief justice called the work of filtering machine-generated arguments unsustainable. [3] [4] [5]

WHY IT MATTERS

the census shows adoption growing faster than client benefit, with checking burdens on both lawyers and clients. Implication: firms serving repeat corporate buyers may face pressure to pass on savings.

Read the full assessment

Demand may also shift toward citation-checked tools and processes for handling clients' AI-generated documents.

Executive brief

AI use among Victorian lawyers jumped from 37% to 55% in a year, but only 26% of AI users say it has lowered what clients pay. Clients' own chatbot use is also making work slower. In the regulator's release, 46% of client-facing lawyers said clients arrive with inaccurate information from AI, and only 10% said clients understand their legal problem better because of it. The data comes from a voluntary 2026 census of 2,221 lawyers run by the Victorian Legal Services Board and Commissioner (VLSB+C). Most lawyers use general chatbots rather than legal-specific tools. The lead researcher said that "widespread use should not be mistaken for widespread benefit."

What changed and event timeline

  1. First Australian AI sanction

    The VLSB+C varied the practising certificate of "Mr Dayal" after he filed AI-generated authorities with false citations. He can no longer run his own practice and must work under supervision for two years.

  2. High Court chief justice warns

    Stephen Gageler said judges had become "human filters" for machine-generated arguments and called this an "unsustainable phase."

  3. Federal Court practice note

    The GPN-AI practice note covers everyone who files documents with the court. Disclosure is required in some circumstances, and non-compliance can bring costs orders.

  4. Baseline report

    The VLSB+C's findings from the 2025 census put AI use at 36.7%. Lawyers mostly framed AI as a productivity tool, not as a way to make services more affordable.

  5. 2026 census findings

    AI use rose to 55% and 78% of users say their use increased. Fee reductions and client-side benefits remain limited.

Capabilities and access

  • No model versions are reported. The survey names consumer chatbots: ChatGPT, Claude, Gemini and Copilot.
  • 79% of AI users use general-purpose tools (VLSB+C).
  • 71% use tools their employer specifies, and 51% use legal-specific tools, up from 46% in 2025 (VLSB+C research centre).
Read the full section
  • No model versions are reported. The survey names consumer chatbots: ChatGPT, Claude, Gemini and Copilot.
  • 79% of AI users use general-purpose tools (VLSB+C).
  • 71% use tools their employer specifies, and 51% use legal-specific tools, up from 46% in 2025 (VLSB+C research centre).
  • The Guardian reports 22% for legal-specific tools "alone." That most likely means lawyers using only legal tools, which would not conflict with the 51% figure (Guardian).

Technical analysis for researchers and developers

  • Survey design: a voluntary, self-reported census of 2,221 lawyers.
  • Measures: fees and workload are lawyers' own perceptions. Nothing was measured from billing data or time records.
  • Comparing years: the 2025 and 2026 censuses ask the same questions, which makes the year-on-year trend useful.
Read the full section
  • Survey design: a voluntary, self-reported census of 2,221 lawyers. Neither summary describes weighting or response rate, so the 55% applies to respondents and is not a confirmed share of the whole profession.
  • Measures: fees and workload are lawyers' own perceptions. Nothing was measured from billing data or time records.
  • Comparing years: the 2025 and 2026 censuses ask the same questions, which makes the year-on-year trend useful. Any change in who chose to respond would still bias it.
  • Implication: the cost of checking AI output is the main thing eating into efficiency gains. That gives a case for tools with grounded, citation-checked output.

Claims and evidence

  • AI use went from 37% to 55%, and 78% of users increased their use (VLSB+C).
  • 26% reported lower fees: 34% with business clients versus 23% with individual clients. 18% took on work that was previously uneconomical (Guardian).
  • Clients' AI use added work for 46% of lawyers and reduced it for 19%. 41% received AI-generated documents that needed correcting (VLSB+C).
Read the full section
  • AI use went from 37% to 55%, and 78% of users increased their use (VLSB+C).
  • 26% reported lower fees: 34% with business clients versus 23% with individual clients. 18% took on work that was previously uneconomical (Guardian).
  • Clients' AI use added work for 46% of lawyers and reduced it for 19%. 41% received AI-generated documents that needed correcting (VLSB+C).
  • The Guardian attributes an 89% verification rate to the 2025 survey (Guardian). The regulator's summaries do not show this figure.
  • No independent dataset corroborates these numbers. All of them come from the regulator's census.

Context and prior work

  • Usage was highest among under-30s in 2025. In 2026 the gap by age has largely closed below age 60 (Guardian).
  • Uptake differs by practice area: 85% in trade practices, but 29% in criminal law and 45% in family law (same source).
  • Courts are also adding controls.
Read the full section
  • Usage was highest among under-30s in 2025. In 2026 the gap by age has largely closed below age 60 (Guardian).
  • Uptake differs by practice area: 85% in trade practices, but 29% in criminal law and 45% in family law (same source).
  • Courts are also adding controls. The Federal Court practice note is one example, and Gageler mentioned a Victorian Law Reform Commission review and practice directions in several jurisdictions (AAP).

Limitations, safety and contested findings

  • Self-selection and self-reporting limit how far the results can be generalised.
  • The two summaries differ on legal-specific tool use (51% vs 22% "alone"), and the difference in definitions is not explained.
  • Community legal centres say clients' AI use has hit them especially hard (VLSB+C research centre).
Read the full section
  • Self-selection and self-reporting limit how far the results can be generalised.
  • The two summaries differ on legal-specific tool use (51% vs 22% "alone"), and the difference in definitions is not explained.
  • Community legal centres say clients' AI use has hit them especially hard (VLSB+C research centre). This matches the regulator's concern that AI could widen existing inequalities (VLSB+C).
  • Rychner says clients' views have not been studied yet (Guardian).
  • Hallucinated citations carry real consequences, including professional sanctions like the Dayal matter.

Business and practitioner implications

  • Pricing pressure follows buyer sophistication. Firms serving business clients should expect to be asked to pass savings on.
  • Fewer than 1 in 10 changed their pricing model. Most gains are being kept inside hourly billing.
  • Clients' AI use is a cost of its own. Firms and legal-aid services may need intake processes for AI-generated documents.
Read the full section
  • Pricing pressure follows buyer sophistication. Rychner links fee cuts to corporate clients who buy legal services repeatedly and use AI themselves. Firms serving business clients should expect to be asked to pass savings on.
  • Fewer than 1 in 10 changed their pricing model. Only 7% adopted fixed fees and 3% unbundled services (VLSB+C research centre). Most gains are being kept inside hourly billing.
  • Clients' AI use is a cost of its own. Firms and legal-aid services may need intake processes for AI-generated documents.
  • For vendors, the reported demand is for verifiable outputs and controls employers can set, not just general chat.
FOLLOW THE EVIDENCE

The source trail.

Sources (7)
A LITTLE LESS NOISE. A LOT MORE CONTEXT.

Stay curious.
Follow the evidence.

Independent perspectives, the original sources, and room for the questions that don't have easy answers.

How we build the brief
Connect with us

Find us where you already read.