Chat GPT, Gamma, Open AI – Oh My!: Ethical Concerns in the Era of Artificial Intelligence
Another day in the office, another motion to draft and file. Time is money, and time is ticking. Artificial intelligence (“AI”) platforms are at your fingertips. You recall seeing how this new adaptive technology can generate information at a high rate of speed. You turn to this technology to produce one of your many motions, and the technology provides a polished, cogent argument with case citations in seconds. You use this argument as the basis of your motion and file it with the court. You are relieved that you have one less task on your plate and one less complex argument to tackle, until an Order to Show Cause why you should not be sanctioned is entered by the court on the basis that the case law in your motion is non-existent. As a mea culpa, you respond to the court’s Order via affidavit acknowledging your use of the AI platform in formulating your motion and that you did not independently verify the authorities provided by the AI platform. This is not a law school hypothetical but an unfortunate cautionary tale that garnered attention when an attorney did just that.
Artificial intelligence, meet the legal profession. Legal profession, meet artificial intelligence. To better understand AI, I looked at Chat GPT and Gamma before tackling this article. In seconds, the AI platforms generated material on complex topics and formatted presentations complete with graphics and authorities. Intriguing as it was, the generated information lacked the voice I imagined for this column. I have been told that the AI technology wants to be interacted with and that I could ask it questions to revise and revamp the information to be more in line with what I was looking for. That is where I drew the line.
How will AI change the ethical considerations for attorneys? If used, should there be a required disclosure at the end of the document? Should the use of AI be incorporated into the Rules Regulating the Florida Bar?
Our legal profession requires practitioners to act in good faith towards each other, towards our clients, and certainly towards the judiciary. The – Rules Regulating the Florida Bar require attorneys to present positions grounded in fact or law, or that could be an extension or modification of existing precedent.[1] At a minimum, this requires using case law that exists and facts that are in the record.
Many emerging AI technologies are geared to providing answers the platforms think you want, regardless of whether they exist. But, these technological parameters are not grounded in the authorities that form the foundations of legal principles. Even in their beta phases, AI platforms such as Chat GPT and Gamma do not have access to the private, well-known legal research engines that have become synonymous with legal research. The AI platforms do, however, have access to any publicly available information prior to September 2011. While it is possible that case law, statutes, and procedural rules produced by these platforms in a legal document could be correct, there is no guarantee that the authorities are correct.
In the legal realm, there are several AI programs that are being tested and are still in beta. These legal-specific programs allow the user to upload documents. Another hurdle arises regarding privilege, inadvertent disclosure of confidential information, claw-back rules, and protective orders since once the information in the document is uploaded, it is now “in the cloud” and can be drawn on for other requests.
AI platforms want to provide the user with a very convincing answer, even if the platform lacks any substantive basis for the specific request. When a question is posed where information is lacking, the platform “hallucinates” an answer. In other words, the answer is a known falsity by the platform which is provided for no other reason than the AI platform’s desire to provide you with a response. Now, replace the AI platform with a lawyer, and a judge as the AI user. Under that altered scenario, an answer or argument that is provided and not formulated on existing law or facts, and is not seeking to change existing law, is a frivolous argument.
Introducing any new product or service requires a period of adaptation. I still recall a popular morning news program discussing “the World Wide Web” and “electronic mail” as foreign concepts. Some may say AI will replace lawyers based on the rapid rate it can generate a document or presentation. AI products will likely boast of the efficiency related to work tasks. Yet, efficiency does not equate to ethics. On its own, AI might not be Icarus; with the unscrupulous attorney, it might be.
Yes, we are at the advent of technological advances when it comes to practicing law. However, AI does not account for ethical considerations or possesses judgment. Attorneys do. Practicing law is more of an art than a science, and “good faith” is more cerebral than linear algorithms that can comb through multiple layers of publicly available information. At least, for now.
Therese A. Savona, Esq., is board certified by the Florida Bar in Appellate Practice and is the current Chair of the Orange County Bar Association Professionalism Committee. She has been a member of the OCBA since 2016.
[1] See R. Regulating Fla. Bar 4-3.1.