VoicePrivate Buy Now

STATISTICS

Lawyers and AI Statistics 2026: Adoption, Ethics Rules and Sanctions

How many lawyers use AI, how much time it saves, what the ethics rules say about client confidentiality, and how often courts sanction made-up citations. Every figure links to its source.

Last updated November 19, 2026. We check each source again when a new edition comes out. Free to cite: please link to this page.

Key figures

How many lawyers use AI

85%of legal professionals report some level of AI adoption at their firms in 2026, up from 79% in 2025 (Clio Legal Trends Report 2026, October 26, 2026)

The American Bar Association asks the question more narrowly: whether the respondent's office is using AI-based technology tools. In its survey run from October to December 2024, 30.2% said yes, and the share rises with firm size (ABA 2024 Artificial Intelligence TechReport, published April 25, 2025).

Firm sizeOffices using AI-based tools
Solo practitioners17.7%
2 to 9 lawyers24.1%
10 to 49 lawyers29.5%
500 or more lawyers47.8%
All respondents30.2%

Source for every row: ABA 2024 Artificial Intelligence TechReport. The same report found that 54.4% named "saving time/increasing efficiency" as the most important benefit, and 74.7% named accuracy as a major concern.

Across legal, tax, risk and government professionals, the share whose organizations were actively using generative AI nearly doubled to 22% in 2025 from 12% in 2024. 41% said they personally use publicly available tools such as ChatGPT, and 17% use industry-specific tools (Thomson Reuters 2025 Generative AI in Professional Services Report, April 15, 2025). These figures cover all of those professions, not lawyers alone.

Billable time and time saved

42%

of a lawyer's workday is projected to be recorded as billable in 2026, up from 39% in 2025 (Clio Legal Trends Report 2026).

3.4 hours

of billable work recorded per day, 53% more than in 2016 (Clio Legal Trends Report 2026).

240 hours

per year legal professionals are expected to free up with AI, up from 200 hours in the 2024 estimate (Thomson Reuters Future of Professionals 2025).

$19,000

average annual value per professional from that time (Thomson Reuters Future of Professionals 2025).

More of the day is being billed, but the rest still goes to admin, intake and writing up. Dictating notes, letters and memos instead of typing them is one of the oldest ways lawyers have cut that time, long before generative AI. Thomson Reuters puts the combined US impact for the legal and tax and accounting sectors at $32 billion a year (Future of Professionals 2025). 95% of professional services respondents expect generative AI to be central to their organization's workflow within five years (Thomson Reuters 2025 Generative AI in Professional Services Report).

What the ethics rules say

On July 29, 2024 the ABA issued Formal Opinion 512, Generative Artificial Intelligence Tools, its first formal opinion on the subject. It applies the existing Model Rules rather than creating new ones. The four parts lawyers ask about most:

  • Competence (Model Rule 1.1). Lawyers "need not become GAI experts" but "must have a reasonable understanding of the capabilities and limitations of the specific GAI technology." Relying on output "without an appropriate degree of independent verification or review" could violate the duty of competence.
  • Confidentiality (Model Rule 1.6). Because many self-learning tools could disclose client information, "a client's informed consent is required prior to inputting information relating to the representation into such a GAI tool."
  • Informed consent. "Merely adding general, boiler-plate provisions to engagement letters purporting to authorize the lawyer to use GAI is not sufficient."
  • Fees (Model Rule 1.5). Lawyers who bill hourly "must bill for their actual time," and "may not charge a client to learn about how to use a GAI tool or service that the lawyer will regularly use for clients."

The opinion also cites a Stanford study finding that leading legal research tools built on generative AI "hallucinate between 17% and 33% of the time" (footnote 14, ABA Formal Opinion 512).

State bar guidance

StateGuidanceDateOn client confidentiality
CaliforniaPractical Guidance for the Use of Generative AI (2026 version, replaces the November 2023 guidance)Approved May 14, 2026An attorney "must not input any confidential information of the client into a generative AI solution that may present material risks to confidentiality or security, absent informed client consent."
FloridaEthics Advisory Opinion 24-1January 19, 2024Recommends getting the "affected client's informed consent prior to utilizing a third-party generative AI program if the utilization would involve the disclosure of any confidential information."
New JerseyPreliminary Guidelines on the Use of Artificial IntelligenceJanuary 24, 2024"A lawyer is responsible to ensure the security of an AI system before entering any non-public client information."
New YorkNYSBA Task Force on Artificial Intelligence reportApril 6, 2024"Lawyers should cautiously use these tools, being mindful of a client's privacy."
TexasProfessional Ethics Committee Opinion 705February 2025"Attorneys must be cautious not to disclose confidential client information inadvertently" when using AI tools.

This page summarizes published guidance. It is not legal advice; check the current rules in your own jurisdiction.

Court sanctions for made-up AI citations

2,312court decisions involving AI-made-up content logged worldwide, 1,571 of them in the US (Damien Charlotin, AI Hallucination Cases database, last updated November 18, 2026, read November 19, 2026)

The best-known early case is Mata v. Avianca, Inc., No. 1:22-cv-01461 (S.D.N.Y.). On June 22, 2023 Judge P. Kevin Castel imposed a $5,000 penalty, jointly and severally, after a brief cited opinions that did not exist, and ordered letters to the judges falsely named as their authors. The database count covers decisions where a court or tribunal dealt with fabricated or false AI-generated content; by its own description it does not track every fake citation ever filed.

The pattern: every rule above asks the same two things. Check what the tool produces, and know where the client's information goes when you use it.

Where client words go when you dictate

Cloud AI tools send the audio or text you give them to the vendor's servers to be processed, which is the point the confidentiality guidance above keeps coming back to. VoicePrivate takes the other route: dictation is turned into text on your own computer, and the audio never leaves it. There is no cloud transcription step to review or consent to for the dictation itself. The Legal Edition adds legal vocabulary and citation formats, and runs on Mac and Windows.

VoicePrivate Legal Edition is $24.99/month or $247/year. Individual plans include a 30-day money-back guarantee: email [email protected] within 30 days of your first payment for a full refund.

Frequently asked questions

What percentage of lawyers use AI in 2026?

Clio's 2026 Legal Trends Report says 85% of legal professionals report some level of AI adoption at their firms, up from 79% in 2025. The ABA's narrower question, whether the office uses AI-based tools, got 30.2% in its 2024 survey.

Do lawyers need client consent to use generative AI?

ABA Formal Opinion 512 says informed consent is required before putting information about the representation into a self-learning generative AI tool, and that boilerplate in an engagement letter is not enough. Florida and California guidance also point to informed consent. Check your own state's rules.

Can lawyers bill for time saved by AI?

Under ABA Formal Opinion 512, lawyers who bill hourly must bill for their actual time, and may not charge a client to learn a tool they will use regularly for clients.

How many lawyers have been sanctioned for AI hallucinations?

Damien Charlotin's AI Hallucination Cases database listed 2,312 decisions worldwide, 1,571 in the US, as of November 18, 2026. Not every listed decision ends in a sanction, and not every case involves a lawyer.

How often is this page updated?

When a source publishes a new edition. The date at the top shows the last real update.

Sources