Consumers lost $8.8 billion to fraud last year, and most do not trust artificial intelligence (AI) chatbots to resolve their online shopping issues. The federal Consumer Financial Protection Bureau (CFPB) reportedly is intensifying its regulation of AI, but with AI-related tools disrupting the fintech industry, consumers do not feel protected. They often resort to illegitimate chargebacks known as friendly fraud, according to Monica Eaton, CEO of Chargebacks911, which has developed a chargeback management solution for merchants. In this Take Five interview, Eaton explains why simple human oversight is imperative when AI is employed in the financial realm.
The CFPB has warned banks about the limitations of chatbots, noting that if the technology is not properly deployed, it provides inaccurate information that compromises customer data and privacy. How are AI and chatbots compromising consumer protection?
“There are actually quite a few issues. So, chatbots are of course becoming smarter and smarter, especially with the advancement of AI. Chatbots are being used for phishing campaigns where consumers can click on an email and then go to a site that navigates them where it looks completely legitimate, and then the chatbot is able to manipulate their information, get more information, and pose as a human. The biggest advantage with ChatGPT is not the AI but it’s actually the natural human language. So, it’s humanized AI, which is amazing because you can’t tell that it came from a computer.
“This technology is rather new, and every bank wants to save money on FTE resources, so they will employ different chatbot systems, and the problem is there’s not enough oversight. Some of the AI in chatbots has even been found to have recorded voices where you can use it for predictive routing for voice responses and voice inquiries. You can also use it online and literally there have been incidences where people’s voices have been completely copied … a voice identity is copied to gain certain banking information and banking apps. So, yes, we’re in kind of a scary age.”
Explain how friendly fraud (illegitimate chargebacks) works and the factors leading to an increase in friendly fraud?
“There’s [often] a lack of due diligence in validating the authenticity of a dispute, so you could have a consumer who bought a pair of shoes from Amazon, and nobody is around to validate whether or not they really got that pair of shoes. And so, they can commit something called friendly fraud by contacting their bank and then claiming that they never received the shoes. Well, now they immediately get a refund and a chargeback is filed, and that is friendly fraud, and I don’t know why it’s called friendly. I guess because of friendly fire.”
How do banks counteract this?
“Whether we’re talking about new incidents of fraud through chatbot technology, which nobody really saw coming, or new incidents of fraud from the least expected criminal, which is actually consumers in many cases, it starts out as something innocent, but bottom line, the common denominator here is the bank doesn’t see it coming. They don’t understand. It’s hard and it’s difficult to detect, so if we look at what banks can do to help prevent chatbot fraud, first and foremost they really need to make sure that they don’t use fully automated chatbots. What that means is they should never completely disengage a human aspect so that they can make sure they understand what’s going on with these chatbots.
“Consumers can do a lot as well to prevent that, so banks can educate consumers and make sure that consumers understand some of the risks and what to look for. In terms of preventing friendly fraud, this just comes down to getting access to more data, reviewing that data, and just being more diligent. We’re in a society where pretty much everything is instant gratification, and I always say we’re in the entitlement era, me included. I totally am an entitled consumer. If I buy something, I expected to be delivered the next day.”
We’re a bit spoiled, in other words.
“Oh, totally spoiled, and you know what? If I buy something online and they say three days, it’s like they just lost my business, so we know what we want. People can’t take candy away from a baby. They’re not going to take things away from us, so as an industry, we need to recognize that all right, now that we have all these spoiled consumers, we’re not going to reduce the volume of complaints, issues, etc., but what we can do is adapt to where the world is growing.
“And so, rather than a bank looking at hey, how can I turn away? How can I tell my customers that they’re lying? No, actually, it’s how can you help them solve the problem in a better way? They’re going to come to you. Actually, don’t fight that and maybe instead of filing a chargeback, which is a negative penalty and a problem for the whole industry, figure out a means to pass that contact detail straight through to the merchant. We’re working on technology like that. There are a lot of other companies doing that, so I guess it’s the whole theme of breaking down the silos and engaging in more collaboration and data-sharing so we can be smarter and more transparent.”
The CFPB has issued a warning, but what regulatory steps is the bureau taking to combat fraud or discrimination based on automated systems in the financial sector?
“There is a lot of discussion. It’s such a huge challenge because people don’t really understand it. On one hand, it’s kind of a Catch-22, a double-edged sword, because there is more regulation in place that says, ‘Hey, you need to listen to consumers. You need to shorten the time it takes to give consumers their money back, and make sure that things are fair because more consumers are buying everything online. More and more consumers are being scammed.’ I mean, there’s all sorts of issues because we’re really transitioning away from cash entirely and almost entirely to cards, at least in the U.S., so with all of that push on decreasing time frames and making things more efficient, here’s what happens — when you make things more efficient, numbers grow. They don’t decrease, so we can’t take friction out of both sides. You have to have an equal balance and that’s our problem today.
“When it comes to trying to add friction and regulate chatbot activity and all these other things, I don’t know how we’re going to do that because we have decided that we’re going to have a frictionless experience for all consumers. So, in order to create that, then we need automation. We need technology. We need smarter ways of interacting because if we don’t have that, we can’t deliver that instantaneous response that every spoiled consumer wants. So, it’s going to take some effort. We just want the perfect utopia. That’s all.”
If you were testifying before Congress, what would you advise them to do if they want to establish more protections?
“We’re not going to change consumer behavior unless we create a consequence for negative behavior. Today, there’s a consequence for merchants and business owners who don’t follow the rules. Well, there’s no consequence for consumers who don’t follow the rules. I remember that I used to write checks. Many of us used to write checks and probably don’t write checks today. However, you go to a grocery store, and you can see there is, at least stateside, there is an NSF fee, or insufficient funds fee. A third party will charge you an additional fee if you happen to write a bad check and get your groceries for free.
“Well, in the online environment, this happens every single day. I can order a product. I can call my bank. I can get a refund, and I got the product for free … Now, there is no consequence in place today. And so, if we look at how checks evolved, it started out as something to replace cash, but eventually it became exploited, probably innocently to start, and then grew to a point where we needed to counterbalance, and there must be a consequence to control this negative behavior. The same thing also needs to happen as we evolve — regulators likewise — to the changing world.
“For example, if I look at chargebacks and disputes, this was created simply as a consumer protection mechanism, but it evolved, rightly so, into a compliance mechanism. Today, it cannot be a compliance mechanism because even Visa has said up to 75% of all chargebacks are invalid. So, how can you use chargebacks as a compliance statistic if even the networks themselves are saying seven out of 10 may be a false indicator. So, the world needs to advance our definitions and the applications of some of this data, and then also create fair consequences for behaviors in order to control the flywheel effect that’s going on in many cases.”
