AI is the hot topic in American law circles. The use of AI in everything from routine searches on the internet, to detailed legal research and drafting, to reducing employee costs and needs makes it broadly appealing to legal researchers and employers alike. However, because the use of AI is such a new, uncharted territory in the law field, the potential for legal issues and liability has not been well thought out. The potential for lawsuits and ethical violations for misuse is high.
AI-powered search relies on a learning model that scrapes info from various websites across the internet and spits out information as if it were an expert in the field. The problem with this model is that some of the information may be inaccurate or just lack the discretion that a professional human would exercise. In the legal field context, AI’s shortcomings have led to lawyers getting sanctioned, companies getting sued, and other problems.
Lawsuits Over AI Giving Bad Advice
Take, for example, the Scott vs OpenAI lawsuit. The lawsuit filed by the family and several branches of Yale Law School alleges that ChatGPT advised the deceased to mix Xanax and Kratom, resulting in the victim’s death. According to the lawsuit, ChatGPT basically gave medical advice to an otherwise “happy, normal kid.” (Kratom is a derivative of the Kratom plant and is sold in many forms for use as a pain reliever and stimulant. It has been the subject of many Kratom wrongful death lawsuits due to its addictiveness and potential for overdose.)
Then there is the Shamblin v OpenAI (ChatGPT) lawsuit alleging that ChatGPT encouraged a kid to kill himself by telling him things like he wasn’t rushing, he was just “ready.” Over in Florida, the Attorney General has recently sued OpenAI, alleging ChatGPT aided and abetted the Florida State mass shooting.
Lawyers Sanctioned over AI Fabricating Cases to Support Legal Briefs and Motions
AI has a tendency to create what it thinks you want, and if the information needed to do this isn’t there, it may just fill in the blanks. This is a serious issue that can get a lawyer in trouble fast. In the law practice, lead attorneys commonly have research assistants, law clerks, or associate attorneys who do the research and draft their briefs and motions for them. It is very easy to become so comfortable with a clerk or associate that you don’t think you need to check their work. If they use AI and you file that document with your signature on it, you are responsible for it, not them. When you sign something and file it with the court, you are representing that you read it and you verify as an officer of the court that it is accurate.” This has resulted in several attorneys filing briefs with false information, faulty arguments, and even completely fabricated cases because they failed to double-check before filing.
In the case of Tom Withers, III vs City of Aberdeen, a Mississippi US District Court Judge disqualified both plaintiffs and defense counsel (4 total lawyers) for filing AI-generated briefs with fake legal authorities cited in their briefs. She barred two of them from appearing in her court for two years! In 2023, A New York judge sanctioned several attorneys $5,000 each for filing briefs with false case citations. In California, an appellate attorney was ordered to pay a $10,000 fine for filing a brief with a false case citation generated by AI.
You would think that when lawyers heard about this happening for the first time, they would be extra careful. However, this has not been the case. A database maintained by the Paris School of Advanced Business Studies contains over 1400 instances of attorneys filing legal documents with fake citations and authorities generated by the imagination of AI!
AI in Legal Advertising Causing Ethical Issues: The Next Problem?
Between AI software/agents created to “automate your marketing” and AI features built into Google, ethical violations caused by AI are yet another unanticipated area of concern for lawyers. For example, Google now has features that allow the advertiser to give Google the latitude to craft your pay per click ads to what it thinks a consumer is likely to respond to. The issue with this is that you are trusting Google’s AI not to create an ad that runs afoul of your State Bar’s advertising rules. More importantly, if you have an advertising company, you are to some degree relying on them not to put you in jeopardy, even though many professional digital marketers have very little knowledge of, if any, of your state bar rules.
In Texas, for example, puffing is a no-no for lawyers. You cannot say I am the “Best” lawyer. Your representations need to be factually verifiable. So you can say you were voted the best X law firm in “x” year by “y” readers’ poll, but you cannot just call yourself the “best lawyer.” But if Google sees people search for that phrase, they may title your ad with that phrase. You also have the issue of approval requirements. In Texas, the State Bar requires that advertisements be submitted for approval to the advertising review committee. If Google changes every advertisement, you cannot possibly submit it.
Ethical Issues Using AI in Law Practice Software
AI is now being used to draft many legal documents other than those filed in court. Medical summaries and chronologies, demand letters, and the appeal of AI is the amount of time it can save. However, AI raises serious ethical concerns when it comes to how lawyers bill for AI in some areas of practice.
Many AI demand packages are billed by the demand they draft or by the file. This is so the attorney has a clear number they can bill back to the client. This may be appropriate in some situations but not others. Take, for example, a billable hour client. Assume it’s $500.00 to run the AI and the lawyer bills at $300 an hour for his time. If an AI demand writer does in 30 minutes what a lawyer needs 4 hours to do, the client is better off paying $500.00 for the software service (assuming it does the job right) than $1200 to the lawyer. So, it makes sense to bill the client $500 for the software. It saved them $700.00.
However, in a contingency fee situation, the lawyer gets paid the same percentage of the recovery regardless of how long it takes the lawyer to do the work. Charging the client $500 for the tool is not to the client’s advantage. It saved the lawyer time, but the client is not paying the lawyer for the amount of time he takes.
If a lawyer is going to use AI, no matter how they bill, I would strongly recommend that they disclose it completely and how it affects the client’s bill in the contract in bold print to be safe. But even with disclosure, there are concerns in the contingency fee space. Can a lawyer disclose this expense explicitly in the contract and justify it that way? A lawyer’s obligation is to act in the client’s best interests over their own. I would struggle to see how saving myself time is in the client’s best interest and not just mine.
Conclusion
In short, while AI can be a powerful, time-saving tool, the technology is still very new and not fully understood. The number of issues that have already been exposed in the legal industry and beyond raises concerns about whether it should be unleashed upon the world without some sort of regulation, monitoring, or restriction. The cases above must serve as a warning to lawyers, parents, and society that anything produced with AI needs to be double-checked, any advice taken with a grain of salt, and any use in business needs to be well-disclosed to the client.
Author
