Do managerial and supervisory lawyers have an ethical duty to supervise AI output?

The Rules of Professional Conduct require that managerial and supervisory attorneys take reasonable steps to ensure that a law firm and its employees comply with the Rules of Professional Conduct and the State Bar Act.

California Rule of Professional Conduct 5.1 provides, in relevant part:

(a) A lawyer who individually or together with other lawyers possesses managerial authority in a law firm, shall make reasonable efforts to ensure that the firm has in effect measures giving reasonable assurance that all lawyers in the firm comply with these rules and the State Bar Act.

(b) A lawyer having direct supervisory authority over another lawyer, whether or not a member or employee of the same law firm, shall make reasonable efforts to ensure that the other lawyer complies with these rules and the State Bar Act.

California Rule of Professional Conduct 5.3 imposes a similar requirement of supervision of nonattorney staff:

(a) A lawyer who individually or together with other lawyers possesses managerial authority in a law firm, shall make reasonable efforts to ensure that the firm has in effect measures giving reasonable assurance that the nonlawyer's conduct is compatible with the professional obligations of the lawyer;

(b) A lawyer having direct supervisory authority over the nonlawyer, whether or not an employee of the same law firm, shall make reasonable efforts to ensure that the person's conduct is compatible with the professional obligations of the lawyer.

What happens if the 'person' generating work product for the law firm is a generative AI application rather than a human? Logic suggests that managerial and supervisory attorneys have the same duty to supervise the work of an AI as they do to supervise the work of the human members of a firm.

Generative AI refers to the use of artificial intelligence to create new content similar to what a human might produce. This technology is based on generative models, which are algorithms that learn to generate new data samples by analyzing patterns in existing data.

A Spring - Summer 2023 Bloomberg Law survey suggested that the legal profession is quickly moving toward a broader understanding and embrace of AI technology. The survey indicated that the two most common use cases reported by lawyers were (1) drafting communications (e.g., drafting emails or letters to opposing counsel) and (2) drafting legal documents.

There have been a number of widely reported instances in which the use (or misuse) of AI has been the subject of judicial criticism. Apparently, the ChatGPT AI platform can "hallucinate" non-existent case citations. Lawyers in New York discovered this to their chagrin when the U.S. District Court for the Southern District of New York sanctioned each attorney $5,000 and ordered them to notify each judge falsely identified as the author of the bogus case rulings cited in an AI created pleading. Other attorneys have discovered the hard way that AI is not only not a supercharged search engine but also has the capacity to "hallucinate."

The incident described above (and others) has caused a number of lawyer organizations to issue guidance regarding the ethics of deploying AI in the practice of law. For example, the ABA has addressed the use of AI in the practice of law in its Resolution 604; the California Lawyer's Association commissioned its Task Force on Artificial Intelligence, which has issued a report; and the State Bar of California's Standing Committee on Professional Responsibility and Conduct has issued its "Practical Guide for Use of Generative AI in Practice of Law."

Language from the ABA Report supporting its Resolution 604 (adopted Feb. 6, 2023) states:

"Responsible individuals and enterprises should be accountable for the consequences caused by their use of AI products, services, systems, and capabilities, including any legally cognizable injury or harm caused by their use, unless they have taken reasonable measures to mitigate against that harm or injury."

The accountability urged by the ABA sounds much like the responsibility imposed on managerial and supervisory attorneys by Rules of Professional Conduct 5.1 and 5.3.

Similarly, in 2023, the State Bar of California's Standing Committee on Professional Responsibility and Conduct ('COPRAC') issued its "Practical Guidance for the Use of Generative Artificial Intelligence in the Practice of Law," which provides, in relevant part:

"A lawyer's professional judgment cannot be delegated to generative AI and remains the lawyer's responsibility at all times. A lawyer should take steps to avoid over-reliance on generative AI to such a degree that it hinders critical attorney analysis fostered by traditional research and writing. For example, a lawyer may supplement any AI-generated research with human-performed research and supplement any AI generated argument with critical, human-performed analysis and review of authorities...

Managerial and supervisory lawyers should establish clear policies regarding the permissible uses of generative AI and make reasonable efforts to ensure that the firm adopts measures that give reasonable assurance that the firm's lawyers and non lawyers' conduct complies with their professional obligations when using generative AI. This includes providing training on the ethical and practical aspects, and pitfalls, of any generative AI use."

Read as a whole, COPRAC's Practical Guidance, like ABA Resolution 604, strongly suggests the responsibility imposed on managerial and supervisory attorneys by RPC's 5.1 and 5.3 applies to the use of and output generated by the use of an AI application in the practice of law.

As managerial and supervisory attorneys are apparently responsible for the use of, and output generated by, an AI application for the benefit of a firm's clients,  they should also consider how to bill clients for work generated by AI.

The Rules of Professional Conduct forbid an attorney from charging or collecting an unconscionable fee. See Rule of Professional Conduct 1.5. It logically follows that charging a client for an AI generated work product as if it had been created by a human being working at an hourly rate would be unconscionable. The costs associated with generative AI may be charged to clients in compliance with the engagement agreement, assuming the engagement agreement explains the basis for all fees and costs, including those associated with the use of generative AI.

Conclusion

As much as luddites among us might wish otherwise, the use of AI in the practice of law is increasingly common. As with other technologies previously adopted into practice, this new technology does not require new ethical rules but does necessitate the thoughtful application of existing rules to this technology. Application of existing rules makes managerial and supervisory attorneys in a law firm responsible for the output of any AI employed by the firm and how the firm charges its clients for that work product.