One topic has come up in conversation after conversation with both clients and prospective clients over the past several months:
“We’re using AI more, but we’re not confident our review process has kept pace.”
It’s a concern we’re hearing with increasing frequency, and for good reason. More than 79% of legal professionals now report using generative AI in some aspect of their work, whether it’s drafting correspondence, summarizing documents, organizing information, or assisting with legal research. The efficiency gains are difficult to ignore, and AI is quickly becoming another tool in the modern law firm’s workflow.
As adoption has accelerated, so have the questions. Firms are no longer asking whether they should use AI. They’re asking how to use it ethically, comply with their professional responsibilities, and build a reliable process for verifying AI-assisted work before it reaches a client or courtroom.
That’s an important shift because while the technology has evolved rapidly, attorneys’ ethical and professional responsibilities have not. AI may help produce the first draft, but the attorney who signs the filing remains responsible for everything in it.
The Headlines Keep Reinforcing the Same Lesson
As AI adoption has accelerated, courts and bar associations have increasingly made clear that these responsibilities are not merely theoretical. Attorneys who fail to adequately review AI-generated work have faced sanctions, while ethics committees and state bars have issued guidance reinforcing the same principle: the duty to verify AI-assisted work remains with the lawyer, not the technology.
Perhaps the best-known example came in Mata v. Avianca, where attorneys submitted a federal court brief containing cases generated by ChatGPT that simply did not exist. The court ultimately sanctioned the attorneys, emphasizing that lawyers, not technology, are responsible for verifying every authority cited in their filings.
That wasn’t an isolated event.
In June 2025, attorneys representing former MyPillow CEO Mike Lindell were sanctioned after filing a brief containing numerous erroneous citations generated through artificial intelligence. The court noted that several cited authorities either did not exist or materially misrepresented the law, despite the attorneys certifying the filing. The sanctions served as another reminder that AI-generated work must always be independently verified.
The Rules Are Becoming More Formal
Courts aren’t the only ones providing guidance.
The State Bar of California recently issued Practical Guidance for the Use of Generative Artificial Intelligence in the Practice of Law, reminding attorneys that AI may assist legal work, but it cannot replace professional judgment. Among other recommendations, the guidance encourages attorneys to:
- Independently verify legal authorities and factual assertions generated by AI.
- Protect confidential client information before entering data into AI platforms.
- Understand the capabilities and limitations of any AI tool they use.
- Maintain independent professional judgment rather than relying on AI output.
- Supervise both staff and technology used in delivering legal services.
California is hardly alone.
The Florida Bar has issued ethics guidance addressing confidentiality, billing, competence, and attorney supervision when using generative AI.
The State Bar of Texas has likewise advised attorneys that existing ethical duties. including competence, confidentiality, and supervision, apply equally when AI tools are involved.
The Pennsylvania Bar Association, New Jersey Supreme Court AI Committee, and several other jurisdictions have released similar guidance or are actively developing formal recommendations. While the language differs from state to state, the message is remarkably consistent. AI changes how legal work is produced. It does not change who remains accountable for it.
The Operational Challenge Most Firms Are Facing
The advice sounds simple enough: verify everything before it goes out the door. In practice, though, that verification takes time. Every citation should be checked against the original authority. Quotations should be compared to the underlying case. Summaries should accurately reflect the court’s holding. Legal authorities should be reviewed to ensure they are still good law and applicable to the issue presented. Factual statements should be confirmed against the source documents. It’s careful, methodical work that can’t be rushed.
The challenge is that most attorneys are already balancing client meetings, hearings, strategy, business development, and billable work. Adding another detailed review process to an already full schedule often leads to one of two outcomes: either the review becomes inconsistent, or attorneys spend valuable time performing work that doesn’t necessarily require attorney-level judgment.
In our experience, that’s less an ethics problem than a capacity problem.
The 6-Point AI Quality Control Checklist
By now, most firms understand they need some level of review for AI-assisted work. The question is what that review should actually include.
Before an AI-generated document leaves the firm, someone should verify that:
- Every cited case exists and remains good law.
- Every quotation accurately reflects the underlying authority.
- Every legal proposition is supported by the cited source.
- Every factual statement has been confirmed against client records or source documents.
- Confidential client information has been handled appropriately throughout the drafting process.
- Exhibits, cross-references, formatting, and document consistency have been reviewed before attorney signoff.
None of these steps are new, but the workflow is. AI-generated work requires firms to build a more deliberate and consistent quality-control process before that work can be relied upon, shared with clients or submitted to court.
The firms adapting most successfully recognize that an effective review process isn’t just a policy, it’s an assigned responsibility. They treat AI-generated work as a first draft, not a finished product, and ensure someone is accountable for verifying citations, reviewing source material, and performing quality control before documents leave the firm.
Who Should Own AI Verification?
Every law firm using AI needs a verification process. The real question isn’t whether that process should exist, it’s who should own it.
When we talk with firms about AI, the conversation rarely starts with technology. It usually starts with bandwidth. Attorneys recognize they need a reliable review process, but many don’t have the internal capacity to dedicate someone to citation verification and quality control every time AI is used.
Some firms expect attorneys to perform every verification task themselves. Others continue to rely on traditional review processes that weren’t designed for AI-generated work. In our experience, neither approach is particularly sustainable as AI becomes a routine part of legal work.
The firms adapting most successfully are building verification into their workflow by assigning it to the right people. Experienced paralegals are uniquely positioned to perform the methodical work that effective AI review requires. They can verify citations, compare quotations against primary authority, confirm factual references, review documents for internal consistency, proofread AI-generated drafts, and identify issues before the document reaches the attorney for final legal review.
That doesn’t replace attorney’s judgment; it strengthens it. Attorneys remain responsible for legal analysis, strategy, and final approval, while experienced paralegals provide a dedicated quality-control layer that helps ensure AI-assisted work has been thoroughly verified before it leaves the office.
For firms without internal capacity, partnering with an experienced virtual paralegal team provides an immediate, plug-and-play verification layer, allowing you to scale your AI output without sacrificing accuracy or adding permanent headcount Rather than asking already busy attorneys to absorb another responsibility, firms can create a repeatable review process that scales alongside their use of AI.
AI Isn’t Going Away
Neither are attorneys’ ethical responsibilities. As AI becomes a routine part of legal practice, we’re seeing more firms focus not on whether they’ll use it, but on how they’ll build reliable processes around it. The firms that succeed won’t necessarily be the ones using the most advanced technology. They’ll be the ones combining that technology with thoughtful oversight, sound professional judgment, and the operational support needed to ensure every document that leaves the firm meets the same standard it always has.
Because AI may generate the first draft, but responsibility still belongs to the attorney who signs the filing.
Is your firm’s review process keeping pace with your technology? Whether you need a dedicated paralegal to handle citation verification or want to strengthen your workflow, we can help. Contact our team today to learn how our virtual paralegals seamlessly integrate into your firm’s AI workflow.
FAQs
Can attorneys rely on AI-generated legal research?
AI can be an effective research aid, but attorneys remain responsible for independently verifying every legal authority, factual statement, and quotation before relying on it in client work or court filings.
What does the California Bar say about generative AI?
The State Bar of California’s Practical Guidance emphasizes that attorneys should maintain competence, protect confidentiality, independently verify AI-generated work, supervise the use of AI, and exercise independent professional judgment at all times.
Are other state bars issuing AI guidance?
Yes. California, Florida, Texas, Pennsylvania, New Jersey, and several other jurisdictions have issued guidance or are actively developing formal recommendations addressing the ethical use of generative AI in legal practice.
What role can paralegals play in AI-assisted workflows?
Experienced paralegals are well positioned to verify citations, confirm quotations, review source documents, proofread filings, and perform quality-control reviews before attorney signoff.
Should firms stop using AI because of these risks?
No. Most ethics guidance recognizes that AI can improve efficiency when used appropriately. The key is implementing consistent review procedures that ensure attorneys maintain responsibility for the accuracy and quality of their work.