iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Law firms can use generative AI to assist with research, drafting, summarization, and other document-heavy work, but the lawyer remains responsible for the work product and the client information involved. The practical test is not whether a tool can produce an answer; it is whether the firm can protect the matter, check the answer, supervise its use, and comply with the rules and client terms that apply.
What AI can do in legal practice
Law firms already use AI-based technology in areas such as technology-assisted review for e-discovery. Generative AI adds tools that produce text and other content in response to prompts. Depending on the task and the controls in place, it can assist with drafting, summarizing documents, organizing information, or supporting legal research.
These are potential uses, not guaranteed productivity gains or substitutes for legal analysis. A polished answer can still be incomplete, inaccurate, or based on a mistaken premise. The lawyer must decide whether the output is fit for its intended use.
Can lawyers use generative AI ethically?
In the United States, the American Bar Association’s Formal Opinion 512, issued in 2024, explains how existing duties under the ABA Model Rules apply when lawyers use generative AI. The opinion is guidance on model rules, not a single rule that automatically governs every lawyer. State adoption, local professional-conduct rules, court orders, client instructions, and engagement terms may differ, so check the authorities applicable to the lawyer, matter, and tribunal.
#1 Best Overall
The ABA’s July 29, 2024 announcement of the opinion says lawyers and firms using generative AI must “fully consider their applicable ethical obligations,” including competent representation, protection of client information, communication, and reasonable fees consistent with time spent. In practice, those duties mean that AI use does not transfer professional judgment or accountability to a tool.
| Professional duty | What it means when using AI |
|---|---|
| Competence | Understand the tool’s relevant capabilities and limits, and assess whether it is suitable for the task. Examine its output for accuracy and completeness. |
| Confidentiality | Understand how the tool handles submitted data, who may access it, how long it is retained, and whether it may be used for further model development. Follow firm-approved tools and matter-level access controls. |
| Communication | Consider the engagement agreement, client instructions, the materiality of the AI use to the representation, and applicable jurisdictional guidance. The cited ABA guidance does not establish a blanket requirement to disclose every AI use to every client. |
| Supervision | Set expectations and provide training for lawyers, staff, and other agents who use AI in client work. A supervising lawyer remains responsible for appropriate oversight. |
| Candor and filings | Verify legal claims, citations, and quotations before relying on them or submitting them to a tribunal. Follow the particular court’s rules and orders. |
| Fees | Keep fees reasonable under the applicable rules and engagement terms. For hourly work, bill actual time spent rather than time a task might have taken without AI; the ABA opinion says the cost of learning a tool generally is not passed on to the client. |
Can client information go into an AI tool?
Only after the firm has established that the specific tool and workflow are appropriate for that information. Do not infer confidentiality protections from a consumer-facing or enterprise-branded interface. Review the provider’s terms and data practices before entering confidential or sensitive material, including retention, access, and any use of submissions to develop or improve models. The available guidance does not settle the terms of any particular provider; those can vary and change.
Rank #2
Firm policy should identify approved tools and permitted data categories, establish who can use each tool, and preserve confidentiality and conflicts controls. A tool’s access design should respect ethical walls and matter-level permissions rather than making restricted information available to users who would not otherwise have access.
Recommended Free Tools
Information to confirm before use
- What data the tool receives, including prompts, uploaded files, and connected-system content.
- Who can access inputs and outputs, and what administrative or matter-level controls are available.
- How long submitted material and generated outputs are retained, and whether deletion is available.
- Whether submitted data may be used for model training or other development.
- Whether contractual terms, client instructions, or applicable law restrict the proposed use.
How should lawyers check AI-generated legal research?
Verify authorities and quotations against authoritative legal sources before relying on them. Check that each cited case or statute exists, that the quotation is accurate, that the authority remains valid, and that it supports the specific proposition for the relevant jurisdiction. A citation that looks plausible is not proof that the source exists or says what the answer claims.
A 2024 empirical study found hallucinations in 17% to 33% of tested outputs from two legal research tools under that study’s particular prompts and benchmark. That range describes the tested outputs, products, and setup; it is not a universal error rate for legal AI or a measurement of every current product. It is a reason to verify results, not a replacement for checking the governing rules or doing the legal analysis.
A practical verification pass
- Trace every authority. Open the cited primary source rather than relying on a generated citation or summary.
- Check the exact proposition. Read enough of the source, including relevant context and procedural posture, to confirm it supports the point being made.
- Confirm jurisdiction and status. Check that the authority applies to the matter and has not been superseded, overruled, or otherwise limited.
- Inspect quotations and factual assertions. Compare quoted language with the original and confirm factual claims against reliable record materials.
- Have a responsible lawyer assess the result. Decide whether the research is complete and appropriate for advice, client delivery, or filing.
Do lawyers have to tell clients when they use AI?
There is no supported blanket conclusion that every AI use must be disclosed to every client. The answer depends on the applicable jurisdictional guidance, engagement terms, client instructions, and whether the use is material to the representation. A firm should resolve those questions before deploying a workflow in a matter, and should not treat silence in a general policy as permission to disregard a client-specific term or instruction.
How widely are legal organizations adopting generative AI?
Thomson Reuters Institute reported that the share of legal organizations incorporating generative AI rose from 14% in 2024 to 26% in 2025. These are publisher-reported survey findings, not a census of all law firms or lawyers. The institute’s April 2025 release describes research surveying nearly 1,800 global professionals across legal, tax, accounting, corporate risk, and government sectors; that full sample should not be described as lawyers or law firms alone.
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Scan for outdated or missing drivers - takes under a minute3Clear out junk files and repair common Windows errorsThe figures indicate growing adoption in the surveyed population, but do not establish that a particular use is safe, compliant, or effective at a given firm. Those outcomes depend on the task, data, tool terms, controls, and review process.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How should a law firm evaluate AI tools?
Compare tools against the firm’s actual matters and workflows, not just demonstrations or broad claims about capability. The sources available for this overview do not establish a current head-to-head vendor feature comparison or specific security certifications. A firm should independently assess its candidates and document the basis for approval.
| Evaluation area | Questions for the firm |
|---|---|
| Legal coverage | Does the tool cover the jurisdictions and authoritative sources the firm needs? Can users identify where an answer came from? |
| Verification | Can lawyers trace citations and inspect source material efficiently? Does the workflow support independent review? |
| Confidentiality and access | What do the contractual and technical controls say about retention, access, submitted data, and matter-level permissions? |
| Workflow fit | Can the firm use it with its document management, research, and case workflows without weakening existing permissions? |
| Supervision and auditability | Can the firm train users, assign review responsibilities, and maintain an appropriate record of use and review? |
| Cost and billing | What is the full cost, and how will the firm bill consistently with the fee arrangement, applicable rules, and client terms? |
A controlled adoption sequence
- Choose a bounded use case. Start with a task whose risks can be managed, and define what a satisfactory result must contain.
- Classify the information. Identify the data involved and state what may not be entered into the tool.
- Review the tool and terms. Confirm confidentiality, retention, access, and any use of submitted data, along with contractual and client restrictions.
- Approve the workflow centrally. Assign an owner, define who may use the tool, establish escalation routes, and specify who reviews outputs.
- Train users. Cover tool limitations, verification, client-data handling, approved use cases, and prohibited uses.
- Require lawyer review. A responsible lawyer should review outputs before they are delivered to clients, used as advice, or filed.
- Monitor and revisit. Track quality, security, cost, and applicable rule changes; reassess approval when a vendor changes its product or terms.
These steps reflect risk-management approaches described by the ABA; they are illustrative, not binding universal standards. A firm should tailor its controls to its jurisdiction, clients, matters, and technology environment.
What remains the lawyer’s responsibility
AI can assist with parts of legal work, but it cannot take responsibility for the advice, filing, or handling of client information. Law firm leaders should approve tools and rules that fit their matters; lawyers should verify work and apply independent judgment; and supervisors should make sure staff understand both the permissions and the limits. Adoption is most defensible when the firm can explain what information entered the system, why the tool was suitable, and how a lawyer checked the result.
Free tools Windows power users keep installed
One-click scans. No signup required.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
