AI Legal Research Ethics: What You Need to Know in 2026
AI legal research raises real professional responsibility questions that aren’t fully settled yet, but the core principles bar associations have converged on are fairly clear. Here’s what you actually need to know.
The duty of competence extends to AI
Most state bars have interpreted the existing duty of technological competence (part of the duty of competence under Model Rule 1.1’s comment) to include understanding the capabilities and limitations of AI tools you use in practice — this isn’t a new, separate ethical duty, it’s the existing competence requirement applied to a new technology. Practically, this means you need to understand that AI can hallucinate citations before you rely on it, not just use it and hope for the best.
Verification is non-negotiable, not optional
Every state bar guidance issued so far agrees on this point: AI-generated legal research and citations must be independently verified before being relied upon or filed with a court. This isn’t a suggestion — multiple attorneys have faced sanctions, and in some cases referrals to disciplinary authorities, for filing briefs with fabricated AI-generated citations they didn’t check. The duty of candor to the tribunal applies fully to AI-assisted work product.
Confidentiality obligations don’t change, but the risk profile does
Your duty of confidentiality to clients applies fully regardless of what tool you use to do research or drafting. The practical risk is that consumer-tier AI products may use conversation data in ways that could implicate confidentiality (depending on the specific product’s data policies), which is why enterprise or business-tier accounts with appropriate data handling agreements are the safer choice for anything involving client information, even non-privileged research questions.
Billing transparency
Some bar guidance has addressed billing specifically: billing a client your standard hourly rate for time that AI substantially compressed, without any disclosure, has been flagged as a potential issue in some jurisdictions’ guidance, though this remains one of the less settled areas. Being transparent with clients about how AI tools factor into your work, and adjusting billing practices accordingly where your jurisdiction’s guidance suggests it, is the safer path.
Disclosure requirements in court filings
An increasing number of federal and state courts have adopted local rules or standing orders requiring either disclosure of AI use in a filing, or certification that all citations have been independently verified. These requirements vary significantly by court and are still evolving — checking the specific standing orders of the court you’re filing in, rather than assuming a uniform national rule, is essential.
Where to find current guidance
Because this is an actively developing area, your state bar’s ethics opinions page and the ABA’s own formal opinions on AI use are the most reliable, current sources — a blog post (including this one) is a starting point for understanding the landscape, not a substitute for checking your specific jurisdiction’s actual current guidance before you rely on AI in a way that could raise a real ethics question.
A Practical Verification Checklist
For every AI-assisted research result, confirm that each case exists, read the actual opinion, check that the quoted language appears in it, and confirm it is still good law using a citator. Record who verified the citations and when, so your file shows the diligence you performed.
Building a Firm AI Policy
A short written policy helps everyone stay consistent. It should list approved tools, forbid entering confidential data into unapproved ones, require human review of all filings, and explain when to disclose AI use to clients or courts. Review the policy periodically because tools and rules change quickly. The American Bar Association provides ongoing guidance on technology and professional responsibility.
Frequently Asked Questions
Do courts require disclosure of AI use? Some judges have standing orders requiring it. Check the rules of each court where you appear.
Who is responsible for AI mistakes? The attorney who signs the filing.
Can junior staff use AI? Yes, with supervision consistent with your duty to supervise.
This article is general information about technology and is not legal advice. Rules on AI use differ by state and jurisdiction, so check your own bar authority.