Tuesday, December 3, 2024

WHEN IT COMES TO COMMERCIAL OFF-THE-SHELF DATABASES OF HIGH-RISK INDIVIDUALS AND ENTITIES, "WE NEED A NEW DRUG"


Before I became World-Check's first (and only) Financial Crime Consultant, I wrote financial crime articles for a competitor, while serving as a compliance officer at a large American company selling seven and eight figure investments. While my readers often commented favorably on my choice of content, they also wondered aloud why the individuals and entities that I wrote about had not as yet made it into the firm's database which they were using to for due diligence purposes. The reason was that my information was more current than what was in the database.

Today, in article by the esteemed Professor Rohan Gunaratna, and noted by the industry visionary David Leppan in his work, AML/CFT compliance tasks, which depend upon databases to vet prospective clients, are often ineffective. The professor points out that generative AI, which lawyers have found to their dismay, can and has created completely bogus and fictional content, just to satisfy one who lodges a query, can be used by bad actors for dark reasons. That's the first problem.

Adding to the conundrum, Mr. Leppan adds in his own work that, due to the logistical problems of keeping large databases current, first-generation databases today are failing in their mission to users, to consistently deliver up-to-date data that is accurate, and can be relied upon.

What is sorely needed is a New Drug, a second-generation AML/CFT database that can: (1) be trusted for accuracy, (2) is current as of today, and (3) will insure that it is free from inaccurate and obsolete content, especially disinformation, whether produced through AI through in the old-fashioned method. Until and unless someone reinvents what is available today, the money launderers will beat the compliance officers every day of the week, and I, of all people, should know. Who will meet that challenge?

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.