USE AI-POWERED AML/CFT PLATFORMS FOR COMPLIANCE, BUT INDEPENDENT AI LEGAL RESEARCH INQUIRIES ARE A DEFINITE NO-NO DUE TO FALSE DATA GENERATED

Back when I was seeking an early termination of the three-year Supervised Release portion of my RICO  sentence, I carefully documented not only the reasons for my sentencing judge to let me off a year early, but the legal authorities in support thereof. Yes, my pro se motion was successful, in large part because the US Probation Office agreed that my post-release activities were solid evidence of rehabilitation.*

If you are a close follower of the cases involving America's disgraced former President, Donald Trump, you may know that his former fixer, Michael Cohen, has, through an attorney, also sought to obtain the same relief from supervision on his criminal case. While in the process of assisting in that matter, Mr. Cohen, who no longer has access to the legal research tools available to him as a practicing attorney, admitted taking a shortcut to obtaining the case law needed in support of his motion, by using an Artificial Intelligence generative AI tool known as GOOGLE BARD to locate the legal authorities required for the motion.

 He did not check the decisions by referring to the actual texts, nor did his attorney. We call that professional negligence where I come from."Attorney's failure to research authorities, constitutes gross negligence," is a phrase I still remember from my first-year law school Research and Writing class. [Attorney and Client, Key Number 44(2)].

Cohen's attorney subsequently included those cases in the filings, which were red-flagged by the judge, because they were non-existent decisions, literally created out of thin air by the AI program. Given the huge amount of publicity last year, when a New York judge publicly fined a lawyer for doing the same thing, with the same horrific results, we wonder why Mr. Cohen chose to engage in such an unprofessional move; perhaps he is not keeping up with legal news to the extent that we are in this blog. Indeed, no less an authority than the Chief Justice of United States Supreme Court this week warned the public of the dangers of using AI programs to conduct legal research, due to the fact that ficticious cases, generated by AI when it fails to find authentic decisions, represent a real danger to unwary users conducting legal research.

Compliance officers conducting due diligence research on new bank clients, or checking out the names of counterparties or entities during transaction monitoring, should NOT be using generic AI products. Use your tried-and-true AML/CFT platforms, and related products in your toolbox that are designed for compliance use. You do not want to find, to your professional and occupational dismay, that your valued AI friend gave you data which does not have a factual basis, meaning that it is fake information dreamed up and created by artificial intelligence, when it couldn't find a fact-based solution. Caveat Compliance;  Ignore Chat GPT, Google Bard and its ilk, when doing your job.