Why the Consumer Financial Protection Agency is concerned about bank bots

Financial Planners

Chatbots are so prevalent in the banking industry, according to the report, that more than a third of Americans (98 million) interacted with a bank chatbot in some way last year. Few enjoyed the experience.

One study found that 8 out of 10 consumers who interacted with a chatbot were “more frustrated,” and that “78% felt less connected to a human after a chatbot failed to meet their needs.” I needed it.” The report also suggests that bots may violate consumer finance laws designed to protect bank customers and cause real harm, such as denying customers the support they need. are doing.

“Poorly implemented chatbots can lead to customer dissatisfaction, loss of trust and even violations of the law,” CFPB Director Rohit Chopra said in a statement accompanying the release of the report.

To date, many financial chatbots have been rules-based, meaning they are programmed with specific, pre-determined responses depending on keywords detected in the user’s question. More complex chatbots are increasingly using natural language processing and machine learning to respond with more human-like and more creative responses.

For banks, chatbots have become a pillar of their customer service teams. The Charlotte, N.C.-based bank said in its latest quarterly report that at Bank of America, conversations with chatbot Erica increased sixfold in three years. In the first quarter of 2020, about 5.3 million customers used his digital assistant to chat with him 27.8 million times. By the first quarter of this year, that number had skyrocketed to his 17.8 million, with bot interactions reaching his 166.7 million.

Betterment, a Flatiron-based automated investing platform, uses chatbots to help its customer service team answer common questions, according to Betterment president Mike Reust.

“But providing good service requires a real commitment to supervision,” Roist added in an email reply. Cranes. “We regularly review our internal data and constantly iterate our tools so that users can fully resolve their issues directly in the chatbot.” He said it would be resolved within the next few days, but live customer service and certified financial planners are always on hand to help.

Over the past six months, many chatbot makers and users have added AI capabilities to their bots. The Times, which has been developing its automated customer service product since 1997, his Square-based Liveperson announced in early May that his users will be able to easily program their own bots to meet customer needs. , launched his AI-powered bot of DIY type. Later that month, Flatiron-based Kasisto, which is developing financial bots used by JPMorgan Chase and TD Bank, cited in a CFPB report, was the first bank designed to address industry-specific needs. It announced what it called a unique large-scale language model. For compliance with extensive rules regarding accuracy, reliability and data privacy.

“Banks can realize significant growth and performance benefits from generative AI, but this requires deep consideration of compliance, accuracy, safety and ethical responsibility,” said Kasisto CEO One Zor Gorelov said in a statement accompanying the May 31 release. That AI product is called his KAI-GPT.

In response to the CFPB’s report, the company added, “We appreciate the CFPB’s continued focus on consumer protection as AI technology is used more extensively across financial services.” The industry is on the brink of major change driven by the power of generative AI, and this power must be applied carefully, responsibly and with governance. ”

JPMorgan Chase & Co. declined to comment on its use of chatbots or the CFPB report.

Source link

Leave a Reply

Your email address will not be published. Required fields are marked *