LLM MODELS POSE SERIOUS RISK TO AML – REGULATORS AND OBLIGED ENTITIES MUST PREPARE

AML vendors are keen to promote the benefits of artificial intelligence (AI) in fighting ML and TF, helping obliged entities detect and block dirty money flows. But every technology can also be wielded by bad actors, and AI is no different. It can be abused to attack and evade KYC controls. Indeed, ChatGPT itself has told MLB how it and its AI cousins might undermine AML/CFT controls. Asked how this might happened, it said attacks might be made via AI assisted: *Synthetic identity fraud generation through merging elements from multiple stolen identities to ...


Full access to this article can be arranged with permission from the client that first ordered it. Please contact us to request access. Entries are uploaded to our archive at least one year after being published by a client – free access is restricted to International News Services journalists for background research only. The article date indicates when copy was filed to a client, not when posted to this archive. Upon client requests, International News Services will remove such articles from the archive or not upload them in the first place. They are included to demonstrate the breadth of topics undertaken by the agency and also to help promote clients’ coverage.