Attorneys using artificial intelligence tools in New York face a growing web of ethical obligations, including contradictory disclosure requirements and the constant risk of relying on inaccurate outputs, according to a recent analysis of the state's legal ethics landscape. Lawyers must understand and verify everything AI produces, safeguard client data, and routinely disclose AI use, or risk violating professional conduct rules.
The pressure comes from multiple directions. Some ethics opinions demand that attorneys disclose AI use to clients, while others treat the technology as just another research tool, creating confusion about when disclosure is mandatory. The consequences of getting it wrong range from procedural sanctions to malpractice claims, especially when AI-generated errors slip into court filings or client advice.
Conflicting disclosure rules create uncertainty
New York's approach to AI disclosure has not been uniform. One set of guidelines may require explicit client consent before using AI on a matter, while another focuses on the lawyer's duty of competence, implying that if the output is verified, no special disclosure is needed. This split leaves practitioners in a bind: over-disclose and risk alarming clients, or under-disclose and face an ethics complaint.
The core issue is the duty of candor and the duty of competence. If a lawyer relies on AI without verifying, they violate the requirement to provide competent representation. If they hide that reliance, they may breach the duty to keep clients reasonably informed. The analysis stresses that transparency is not optional-it is a baseline expectation.
A six-step protocol for ethical AI use
The analysis outlines a practical framework to help attorneys stay compliant. The six steps are: knowing the AI tool, securing data, verifying work, disclosure, documentation, and responsibility. Each step addresses a specific ethical risk.
- Know the AI tool: Understand how the tool works, its limitations, and the data it was trained on.
- Secure data: Confirm that client information stays confidential and is not used to train third-party models.
- Verify work: Check every output for accuracy, relevance, and legal soundness before using it.
- Disclosure: Decide what to tell the client about AI use, and document that decision.
- Documentation: Keep records of AI interactions, including prompts and outputs, to support the work product.
- Responsibility: Accept that the lawyer, not the software, answers for the final work product.
Verification remains the most critical step. AI can fabricate cases, misinterpret statutes, or produce plausible but wrong summaries. A lawyer who files a brief without checking citations risks the same fate as those already sanctioned for citing nonexistent cases generated by AI. Rigorous review is the only defense.
Data protection and the duty of confidentiality
Client data fed into AI tools can trigger a confidentiality breach if the platform uses inputs to train its models. Attorneys must review terms of service, choose enterprise-grade tools with data isolation, and avoid free consumer products that lack legal-grade protections. The obligation to protect client secrets extends to every digital tool a firm uses.
Why this matters for legal professionals
Ethical AI use is not a future concern-it is a present compliance requirement. New York attorneys who treat AI like a search engine put their licenses and clients at risk. The six-step protocol gives a concrete checklist, but it only works if lawyers actually follow it, especially the verification step. Firms that fail to build these habits now will face sharper scrutiny as courts and disciplinary bodies harden their stances on AI-related misconduct.
Your membership also unlocks:
