Risk Update

Law Firm Agentic AI Ethics & Risk — New State Ethics Opinion and Guide, BRB Survey Reminder

Posted Dan Bressler

David Kluft asks: “What is Agentic AI and do any state ethics opinions talk about it?” —

  • “Agentic Artificial Intelligence (AAI) has been defined as ‘AI systems that autonomously plan and execute multi-step tasks with minimal human oversight, interacting with software tools and adapting in real-time.'”
  • “The AL bar recently issued a guide to best practices for lawyer use of AI. For the most part, it goes over well-worn ground that you can find in other opinions (hallucinations, confidentiality, etc.), but it also has a section on ‘Specific Agentic AI Concerns.'”
  • “It concludes that ‘AI is not so reliable that embedding core values alone will ensure compliance or prevent harm,’ and therefore ‘attorney oversight will always be critical, regardless of attempts to program AI with governing principles.’ The opinion adopts a set of best practices for Agentic AI (AAI), including:
    • • Do not deploy AAI tools on client matters without a documented human review checkpoint before any action is taken on the AI’s behalf. “
    • • Define in writing what tasks AAI is and is not authorized to perform autonomously. ”
      Audit AAI activity logs regularly to identify errors, unauthorized actions, or outputs that require correction. ”
      Treat AAI governance as you would any other high-risk workflow requiring documented protocols and supervisory sign-off. ”
      Consult with AI vendors about what safeguards and audit trails are built into their agentic AI products before deploying them.”
  • Read the guide: here.

The California Bar has its own guidelines: “PRACTICAL GUIDANCE FOR THE USE OF GENERATIVE ARTIFICIAL INTELLIGENCE IN THE PRACTICE OF LAW” —

  • “Generative AI products have increasingly incorporated ‘agentic’ capabilities, which can enable systems to autonomously perform tasks or workflows without human prompting. Agentic systems are built not just with multistep functionality but also can pursue goals, ‘plan,’ select tools, and even execute tasks without immediate human intervention. Examples include pleading revisions across multiple iterations, preparing discovery responses, coordinating review of documents, and autonomously facilitating client intake.”
  • “Agentic AI introduces an additional degree of autonomy, including the ability to initiate tasks or interact with external tools, and the capacity to conclude tasks without human review or approval. Yet, this autonomy does not satisfy a lawyer’s duty to exercise independent judgment, nor does it alter a lawyer’s existing ethical obligations. Rather, the ability of agentic AI systems to initiate tasks or interact with external tools increases the risk that a lawyer may rely on automated processes without contemporaneous review.”
  • “Agentic AI systems may be configured to access internal firm systems, such as email, messaging platforms, document management systems, knowledge bases, client files, or calendaring systems. Unlike isolated prompt-based tools, such systems may have persistent or automated access to large volumes of confidential client information, raising acute confidentiality considerations under rule 1.6 and Business and Professions Code section 6068(e)(1). “
  • “Lawyers must carefully evaluate and limit the scope of such access. Unrestricted or poorly configured agentic systems may unintentionally disclose confidential information (including across different matters) and even expose privileged material. As a result, a lawyer remains responsible for ensuring that any agentic AI system is configured in a manner consistent with their duty of confidentiality and that appropriate monitoring and access controls are implemented. A lawyer must not deploy an agentic AI system in a manner that permits autonomous external transmission of client information, including automated communications, filings, or data transfers, without appropriate safeguards and human review.”
  • “Managerial and supervisory lawyers should establish clear policies regarding the uses of generative AI, including more autonomous or agentic AI tools, and make reasonable efforts to provide reasonable assurance that the conduct of the lawyers and nonlawyers under supervision complies with the lawyers’ professional obligations when using such technologies. This includes providing appropriate training on the ethical and practical aspects, and pitfalls, of any AI-enabled tools that perform tasks with limited or no real-time human direction.”

See also: Oregon FORMAL OPINION NO 2026-208.

 

And please do consider participating in the Risk Blog’s Agentic AI survey. It’s a great way to support my work here at the blog! Read more: here.