The full weight of the AI industry is leaning on the legal profession. The large legal research providers to smaller niche companies have unleashed a constant and unrelenting barrage of marketing aimed at lawyers. They intend for us to relinquish time and professional discretion in the conduct of core lawyer activities.
Many have succumbed, ceding to algorithms the tasks of preparing substantive letters or emails, and legal memorandums for submission to courts, the latter often producing dire consequences for both lawyers and clients.
I have always been an early adopter of tech. My legal career started shortly before the birth of desk tops, portables, and all things Microsoft. My professional development has followed an arc roughly congruent with introduction of DOS, Windows, portables, cell phones, the internet, Google, email, practice management, client data management, cloud computing and social media. It has been a wild ride, but none of this sought to replace the core function of your lawyer-- contemplative thought and problem solving. In fact, almost all tech developed to date freed your counsel to shed administrative tasks in favor of deeper considerations.
Courts across the Country are now admonishing and punishing the use of "hallucinated cases" in briefs and argument. Large firms and small, alike, are falling to AI industry pressure to cede research and writing to tech. The lucky perpetrators survive with warnings and fines. The less fortunate lose the privilege to practice law, and their clients suffer catastrophic loss in their cases.
One recent decision out of Georgia demonstrated the harm of judges relying on hallucinated case citations copied from the court submissions of lawyers in the case. It was a situation where a trial judge ratified the fake legal conclusion of a fake legal precedent submitted by a lawyer who was too lazy to double-check the product generated by his AI product. Imagine the thousands of wasted dollars expended by the parties to unwind this fiasco through the appeal.
In another jurisdiction, a judge caught the use of hallucinated case law by one lawyer, only to admonish opposing counsel for their failure to bring the situation to the court's attention. This signals a trend to impose additional burdens on your lawyer to actively seek out and catch your opponent's reliance on fake law. At the very least, it might cost ten minutes of billable time to check a single case citation in a brief. It would add great expense for your lawyer to now review every single citation, even for standard propositions that often need not be double-checked, for fear of violating the newly imposed duty to back check opposing counsel's use of AI.
Very recently, I was being interviewed by a prospective client over the telephone. In real time, during our call, I was being told that "well, ChatGPT says you are wrong..." I simply asked the prospective client "does ChatGPT have malpractice insurance, if it is wrong?"
Within the last month, prospective clients have sought to engage our Firm to fix their botched AI court submission, and to avoid dismissal of their cases by really angry judges, and to avoid sanctions sought by really angry opposing lawyers. At much additional cost to the clients, we can do this, and only then return the case to a consideration of the merits. Every dollar spent to return the case to normal is both necessary and wasted.
One prospective client presented our Firm with their "completed" motion and memorandum, comprising over thirty-pages and hundreds of court citations. They were surprised and disappointed that I would decline such a "slam dunk" where all the work has been completed by AI, and I needed to only sign my name and file with the court. As I scanned the proposed submission, I could hear the snakes slithering in and through the woodpile of AI generated case citations, none of which I recognized after almost four decades of practice.
A very significant risk created by over-use of AI was reported, this past week, by a New York judge. The judge held that use of AI, which includes typing in very specific facts and details about one's case, can operate to obliterate the protection of the attorney-client privilege. This means that otherwise privileged information, which need not be shared with the opposition in discovery, becomes exposed to mandatory disclosure because it was already shared with a third-party, the AI platform and algorithm. Over the immediate short term, this will be the hottest area of litigation, and the greatest multiplier in your litigation cost-- the fight over whether your use of AI, or the use by your lawyer of AI, in the conduct of your case has stripped away the ordinary protections of the attorney-client and work-product privileges.
AI has added another wrinkle to how you hire and work with lawyers. Please insist that lawyers you hire conduct their own legal research. AI can certainly improve the prose of otherwise poor writers, just as it can ruin the product of otherwise polished brief writers. But you should never accept the risk that your lawyer has ceded their legal research and analysis function to AI. It should be something that is written in your fee agreement or engagement with the lawyer or Firm.
