Experian uncovers fraud paradox in financial services’ AI adoption

Host A: Welcome back to AI Catchup Weekly, I'm here with my co-host, and we've got a story today that is equal parts fascinating and genuinely unsettling — Experian just dropped their 2026 Future of Fraud Forecast, and the headline finding is basically this: the AI that's supposed to protect us is also being used to attack us.

Host B: Right, and that's not just a clever soundbite — there are real numbers behind it. What are we actually talking about in terms of scale here?

Host A: So according to FTC data cited in the report, consumers lost more than twelve and a half billion dollars to fraud in 2024. And nearly sixty percent of companies saw their fraud losses increase from 2024 to 2025. On the flip side, Experian says their own fraud prevention tools helped clients avoid an estimated nineteen billion dollars in losses globally last year.

Host B: Nineteen billion. That's a staggering number, and it really puts into perspective how much of this fight is now AI versus AI — it's not humans catching fraudsters anymore, it's algorithms racing against other algorithms.

Host A: Which brings us to what Experian calls the most pressing issue in the whole report — they're calling it "machine-to-machine mayhem." The idea is that agentic AI, systems designed to browse, transact, and make decisions on your behalf, are becoming indistinguishable from the bots fraudsters are using to do the exact same thing.

Host B: And that raises a genuinely tricky question, right? If an AI agent makes a fraudulent transaction — maybe it's been compromised or spoofed somehow — who's actually liable? The person who deployed the agent? The company that built it?

Host A: Exactly, and that's the uncomfortable truth — there's no settled answer right now. Experian predicts 2026 is the year that forces a real industry-wide conversation about governance and liability. Amazon has already pre-emptively said it's blocking third-party AI agents from browsing and transacting on its platform entirely.

Host B: Smart move, honestly. But beyond the agentic AI issue, there are four other trends in this forecast that caught my eye — and some of them are genuinely creepy. Deepfake job candidates?

Host A: Yeah, this one is wild. Generative AI can now produce fake CVs and real-time deepfake video convincing enough to pass a job interview. The FBI and Department of Justice both issued warnings in 2025 about North Korean operatives literally using this technique to get hired at US companies and gain access to internal systems.

Host B: So the threat isn't just at the consumer level — it's inside organizations. And then there's the emotionally intelligent scam bots, which honestly sound like something out of a sci-fi movie.

Host A: These are bots that can run romance fraud and "relative in need" scams entirely without a human operator — they build trust over weeks or months, respond naturally, and are increasingly hard to distinguish from a real person. The fact that this is now automated and scalable is a really dark development.

Host B: And then on top of all that, your smart fridge might be a security risk. I'm only half joking — smart home devices are flagged as a new attack vector too.

Host A: Completely true. Virtual assistants, smart locks, connected appliances — Experian says bad actors are looking to exploit all of it to access personal data and monitor household behavior. So how are financial institutions actually responding to all of this?

Host B: That's what I want to know, because awareness is one thing — actually doing something about it is another.

Host A: So Experian surveyed over two hundred decision-makers at financial institutions, and eighty-four percent say AI is a critical priority for their strategy over the next two years. But sixty-five percent admit that AI-ready data is one of their biggest challenges, and sixty percent are still running manual compliance processes. There's a real gap between ambition and execution.

Host B: And that data quality point keeps coming up across the industry — it's not just Experian saying it. Your AI is only as good as the data it's trained on, and in financial services where every decision needs to be explainable and auditable, that's not a small problem to paper over.

Host A: Exactly right. It's the unglamorous foundation that everything else depends on. Alright, that's going to do it for today's deep dive — fraud is evolving faster than most of us realize, and 2026 is shaping up to be a genuinely pivotal year for how the industry responds.

Host B: If nothing else, maybe think twice before your AI assistant books a flight on your behalf. Thanks for listening to AI Catchup Weekly, we'll see you next time.