Will AI Replace Lawyers? What AI Really Does in Law Firms in 2026

No, AI will not replace lawyers, but it is already changing how they work. Goldman Sachs estimates that 44% of legal tasks could be automated, yet the Bureau of Labor Statistics still projects lawyer jobs to grow 5% from 2025 to 2035. Research, drafting, and document review can be automated; judgment and professional responsibility cannot.

In this article, you will see what AI can really do in a US law firm today, how ABA Formal Opinion 512 frames its use without banning it, and which tools to choose so that AI becomes a competitive edge rather than an ethics risk.

AI passed the bar exam: what does that really mean?

The argument always comes back to the famous exam. In a March 2023 experiment led by researchers from Chicago-Kent College of Law and the legal tech company Casetext, GPT-4 scored about 297 on the Uniform Bar Exam, above the passing score in every UBE jurisdiction. Be careful, though: the famous "90th percentile" estimate has since been seriously challenged. An academic re-evaluation by MIT researcher Eric Martínez places GPT-4 below the 69th percentile of all test takers, because the original comparison leaned on repeat takers, who score lower.

Impressive, but it needs serious context. So can AI replace lawyers on the strength of an exam? Testing legal knowledge is not the same as practicing law, and benchmarks on real legal questions show the gap: in a 2024 benchmark by Stanford RegLab and HAI, general-purpose chatbots hallucinated on 58% to 82% of legal queries, and even legal research tools got it wrong more than 17% of the time.

The National Conference of Bar Examiners, which designs the multiple-choice section of the exam, also pushed back: lawyers have unique skills, gained through education and experience, that "AI cannot currently match."

What AI already does (very well) in US law firms

Adoption is no longer a forecast; it is a daily reality. According to Clio's 2026 Legal Trends Report for Solo and Small Law Firms, based on a survey of 1,702 US legal professionals, 71% of solo practitioners and 75% of small firms now use AI in their work. Going further pays off: in Clio's 2025 Legal Trends Report, firms with wide AI adoption were nearly three times more likely to report revenue growth than firms that had not adopted AI.

In practice, AI excels at tasks that take time but follow a clear frame. The most mature use cases today:

• Case law research in plain English
• Contract review and comparison, with risky clauses flagged
• First drafts of pleadings, memos, and demand letters
• Document summaries and case chronologies
• Qualifying and sorting incoming inquiries

This is exactly what the Goldman Sachs estimate measures: tasks, not jobs. The report describes what AI can do in this area as reviewing documents for compliance with legal, regulatory, and corporate standards, work that sits squarely on this list.

On the last item, qualifying inquiries, AI meets marketing: set up well, it helps qualify prospective clients and smooths the first contact. It is a lever we work on as part of client development. If consumer tools are what interest you, we have detailed what you can (and cannot) do with ChatGPT at your firm, and our guide to AI for lawyers walks through every practical use.

What AI cannot do

There is a clear line, and it comes from the nature of the job itself. Legal AI is an assistant that speeds up research, document analysis, and first drafts. It does not replace legal reasoning applied to a new situation, the careful characterization of the facts, case strategy, or the relationship of trust with a client.

There is also a major technical risk: hallucination. No legal AI tool replaces the lawyer, and every one of them, without exception, still produces invented content. The problem has reached the courtroom. The database maintained by researcher Damien Charlotin, which tracks court decisions involving AI-fabricated citations, listed 2,199 cases worldwide as of October 16, 2026, including 1,501 in US courts.

The best known remains Mata v. Avianca. In June 2023, a federal judge in New York fined two lawyers and their firm $5,000 for filing cases that ChatGPT had made up. As the opinion puts it, there is "nothing inherently improper about using a reliable artificial intelligence tool for assistance," but the rules give attorneys a gatekeeping role over the accuracy of their filings. A lawyer who files a fake case is responsible, period. That is precisely what the machine will never carry for you.

ABA ethics rules already govern AI use

This is the central point, and too few articles treat it seriously. The profession did not wait. On July 29, 2024, the American Bar Association issued Formal Opinion 512, its first formal ethics opinion on lawyers' use of generative AI tools. It builds on comment 8 to Model Rule 1.1, which since 2012 has asked lawyers to keep up with the benefits and risks of relevant technology, and several state bars have added their own guidance.

The philosophy is clear and reassuring: no ban, but a requirement of control. The opinion does not prohibit AI. It allows it as long as the lawyer meets every existing duty and stays fully responsible for the work. Opinion 512 organizes its guidance around core obligations: competence, confidentiality, communication with clients, candor toward the tribunal, supervision, and reasonable fees. It is explicit that relying on AI output without an appropriate degree of independent verification can breach the duty of competence.

The strictest rule concerns your client data. Rule 1.6 protects all information relating to the representation, including when you use an AI tool. Before you enter client information into a self-learning tool, Opinion 512 says you need the client's informed consent, and a boilerplate clause in your engagement letter is not enough.

The simplest safeguard is to anonymize data systematically before you submit it, replacing identifying details (names, addresses, matter numbers) with neutral codes. This discipline matches what we recommend on any sensitive topic, such as the solicitation rules: ethics is not a brake, it is a framework.

Client confidentiality: why not all AI tools are equal

Choosing a tool is not just a matter of comfort; it is a matter of compliance. The dividing line runs through the terms of use. In United States v. Heppner, decided in February 2026, Judge Jed Rakoff of the Southern District of New York held that a defendant's exchanges with the consumer version of Claude were protected by neither attorney-client privilege nor the work product doctrine.

One reason: the provider's privacy policy allowed it to use inputs for model training and to disclose data to third parties, so there was no reasonable expectation of confidentiality. The defendant used the tool on his own, not at his lawyers' direction, but the lesson for firms is the same. Venable's analysis of the ruling shows how quickly consumer terms can undermine the confidentiality you owe your clients.

Opinion 512 does not name brands, but it gives precise criteria: read the terms of use, privacy policy, and contractual terms of any tool to learn how it collects and uses what you enter, or rely on someone with the expertise to analyze them. Before you sign, require a written commitment that your data will not be used to train models, not just a line in the terms of service.

State privacy laws such as the CCPA may also apply, and GDPR comes into play if you handle data from clients in Europe. These compliance habits also apply to how you build your online presence, starting with your law firm website.

AI is a competitive lever more than a threat

Let's flip the perspective. Firms that adopt AI intelligently win back time where they used to lose it: in Thomson Reuters' 2025 Future of Professionals report, professionals expected AI to save them about five hours a week, or roughly 240 hours a year. The real question then becomes strategic: what will you do with that freed-up time?

Will lawyers be replaced by AI? The job market says no, and points to a future of AI in law built on augmentation. Despite the automation estimates, the BLS expects employment of lawyers to grow faster than the average for all occupations through 2035, with about 28,700 openings a year. Lawyers are not disappearing; the ones who use AI well are pulling ahead.

The real threat is not the machine; it is the governance gap. In Clio's 2026 report, 57% of solo firms and 55% of small firms have no AI policy in place, compared with 24% of enterprise firms, and only about a third of solos (32%) and small firms (31%) have increased revenue with AI. That is exactly where the competitive edge lies: mastering the tool and reinvesting the time saved in high-value advice, client relationships, and your firm's visibility.

On that last point, AI is also changing client acquisition, with the rise of AI search optimization, where being cited by AI assistants becomes a channel of its own, alongside a solid SEO strategy.

Where to start: 3 tools built for US legal practice

No need to aim for the solutions large firms use: Harvey targets large law firms and in-house legal teams, and it remains out of reach for most small practices. For most of the market, here are three accessible options:

A good habit before paying: check with your state bar. Many offer free legal research as a member benefit, as the State Bar of Arizona and the State Bar of Texas do with vLex Fastcase. Whatever the tool, keep the golden rule: every output must be checked against primary sources before it goes into a professional document. For a wider comparison, see our roundup of legal AI tools.

Adopting AI is a decision about organization and strategy, just like your choices about visibility. To go further, our guide to law firm marketing lays the foundations for a firm that combines modern technology with controlled growth.

Key takeaways

AI is not a competitor to lawyers but an assistant that speeds up well-defined tasks while leaving advice, strategy, and responsibility intact. ABA ethics rules allow it on three conditions: control, client confidentiality, and systematic verification of sources. The real challenge is not to resist, but to choose tools with sound data terms and reinvest the time saved in what makes you valuable.

AI will not replace lawyers, but lawyers who master AI will pull ahead of those who ignore it. If you want to build a digital and AI strategy that fits your ethics obligations, our marketing agency for lawyers is here for exactly that. You can also tell us about your firm directly.