Banks Warned Against Relying on Error-Prone, Alienating Chatbots

Banks’ growing use of AI-powered chatbots risks alienating customers by providing incorrect information and reducing human contact, the Consumer Financial Protection Bureau said.

Banks in recent years have deployed increasingly sophisticated online chatbots to quickly answer customer questions while cutting back on human customer service agents. They’ve also been experimenting with the use of generative machine-learning tools such as OpenAI’s ChatGPT and Alphabet Inc.’s LaMDA service.

But those tools are often unable to answer complex questions from consumers, and in some cases can give people the wrong answer, the CFPB said in a Tuesday report.

Some chatbots have been ...