PROFESSIONALISM CORNER
want to attend our next cle or upcoming event ?
PROFESSIONALISM CORNER
Generative Artificial Intelligence — Master or Servant ?
D . CULVER " SKIP " SMITH III
Look around , it ’ s all so clear , Wherever we ’ re goin ’, well it ’ s here .
— Paisley , Brad . Welcome to the
Future [ song ].
Unless you have been hibernating the last year or so , you know that the current hot topic for lawyers is how generative artificial intelligence is finding its way into the practice of law . Its potential uses and efficiencies are nearly boundless . Pitfalls also abound . You likely have read about the New York lawyer who was sanctioned for filing a memorandum created by generative AI that cited nonexistent judicial opinions and contained fake quotes . See Mata v . Aviance , Inc ., 2023 WL 4114965 , 2023 U . S . Dist . LEXIS 108263 ( S . D . N . Y . June 22 , 2023 )
The fact is that generative AI can “ hallucinate ” or create “ inaccurate answers that sound convincing .” Matt Reynolds , “ vLex Releases New Generative AI Legal Assistant , A . B . A . J ., https :// www . abajournal . com / web / article / vlex-releasesnew-generative-ai-legal-assistant ( Oct . 17 , 2023 ). Some judges have overreacted to generative AI , requiring disclosure of its use or in one instance even banning the use of all AI ( not just the generative variety ). These reactions are overbroad : they fail to recognize that AI for quite some time has been with us in more innocuous forms , such as cloud-computing . Microsoft ’ s autocomplete function ( used in the writing of this article ) is a form of generative AI . They also are unnecessary : court rules and ethics rules amply address improper uses of generative AI . ( The Mata court acknowledged this .)
The Florida Bar recently issued Ethics Opinion 24-1 regarding the use of generative AI . Fla . Bar Eth . Op . 24-1 , 2024 WL 271230 , ( Jan . 19 , 2024 ) ( also available at https :// www . floridabar . org / etopinions / opinion-24-1 /). The opinion addresses several potential ethical issues , including confidentiality , oversight ( verification of accuracy ), billing practices , and advertising .
The opinion concludes that a lawyer may ethically utilize generative AI technologies but only to the extent that the lawyer can reasonably guarantee compliance with the lawyer ' s ethical obligations .
The principal concerns are confidentiality and oversight :
Confidentiality . Generative AI does not raise novel confidentiality issues any more than did cloud-computing and online storage . Although the risk of unwanted disclosure of confidential information likely is low , Ethics Opinion 24-1 recommends obtaining client consent to the use of generative AI . Whether this is advisable is debatable . As the opinion itself acknowledges , Rule Regulating the Florida Bar 4-1.6 ( c ) ( 1 ) permits disclosure of information relating to the representation when reasonably necessary to “ serve the client ’ s interest unless it is information the client specifically requires not to be disclosed .” The lawyer ’ s use of generative AI serves the client ’ s interest . Seeking the client ’ s consent to its use ( via the engagement agreement or otherwise ) is unnecessary . It also suggests to the client that withholding consent might be desirable , which would create a practical nightmare for the lawyer . ( The same holds true for cloud-computing and cloud-based storage .)
Oversight . The New York lawyer ’ s experience in the Mata case is a lesson for all : check the accuracy of the product . Simple fact-checking is not enough . Generative AI not only is susceptible to inaccuracies , but also is biased , because it is influenced by the data that it has received over time . Thoughtful review by the lawyer is essential , not just whether the product is accurate , but also , “ Does this seem right ? Does this make sense ?”
Much of this is common sense , but common sense sometimes seems less common these days .
The use of generative AI doubtless will continue to expand in the practice of law . It will become a frequent tool of judges as well . ( Judge to J . A .: “ Load the briefs into the AI program , and let ’ s see what ruling it produces .”) And why shouldn ’ t it ? This is not to suggest that generative AI someday will replace judges ( or lawyers ). So much in the application of law to facts is nuanced , and human judgment is essential to much of what judges and lawyers do every day . Generative AI programs derive their capabilities from absorbing large sets of data containing content created by humans . To create a product , generative AI requires prompts , feedback , and guidance from humans . The quality of the content that it creates depends on how well thought out the prompts , feedback , and guidance are . Human intelligence — and judgment — remain essential .
Anything that helps produce the best possible product is worthy of using — so long as it is used as a tool for achieving a result and is not the last word . Proceed — but with caution . A lawyer must treat generative AI as one ’ s servant , not one ’ s master .
Welcome to the future .
( A word of encouragement : Don ’ t deny yourself the use of this amazing tool . When the author of this article began practicing law , carbon paper and Wite-Out were tools of the trade . The first big new thing was the IBM Selectric typewriter . ( A roller and typeball — imagine !) Personal computers were years away . Dictaphones ruled . Smartphones ? Still , the author has managed , lo these many years , to keep up with — indeed , happily embrace — most new technologies as they arrived on the scene . You can do it !)
want to attend our next cle or upcoming event ?
view our full calendar at palmbeachbar . org / calendar
PBCBA BAR BULLETIN 15