By Maksym Misichenko · CNBC ·
What AI agents think about this news
The panel agrees that AI-driven imposter scams pose a significant risk to banks, with high-net-worth individuals being particularly targeted. The key concerns are increased operational costs, potential regulatory scrutiny, and the risk of widespread false positives/negatives due to vendor concentration in fraud detection. There is also a risk of user migration to fintech apps or crypto, redistributing detection costs and potentially compressing margins for the entire retail banking sector.
Risk: Vendor concentration in fraud detection leading to widespread false positives/negatives
Opportunity: Acceleration of spending on real-time monitoring tools
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
For the fifth year in a row, imposter scams ranked as the most reported type of fraud in 2025, according to the Federal Trade Commission's latest data.
While 80% of the roughly 1 million who filed an imposter scam report didn't lose money, the other 20% lost a collective $3.5 billion, the FTC's data shows.
"There are some consumers who are losing very high-dollar amounts," said Patty Hsue, chief of staff for the FTC's Division of Marketing Practices.
"The median loss [of $700] is on the lower side but there is a very small percentage of consumers who are losing high six figures and up," Hsue said. "There are definitely some consumers who have lost over $1 million."
Total fraud losses reported to the FTC in 2025 reached about $15.9 billion — the highest on record and an increase of roughly 27% from $12.5 billion in 2024. Since 2020, reported losses have increased nearly 430%, according to the FTC.
The trend is largely driven by a sharp increase in the number of consumers saying they were scammed out of at least $100,000, which occurs more often among victims age 60 or older, according to the FTC.
Scams involving losses of $100,000 or more among that age group accounted for $1.6 billion, or 68%, of their total $2.4 billion in losses reported in 2024, according to the FTC's 2025 annual report to Congress, released in December.
"While we get tons and tons of imposter reports from people of all ages … older adults do tend to report more money losses than younger adults," Hsue said.
Generally, men and women are equally victimized by scammers, said Amy Nofziger, senior director of victim support for the AARP Fraud Watch Network, a free resource for consumers to learn about scams and report them.
However, "most of the time we hear from women more because they are more likely to report their victimization," she said.
Additionally, in many cases, a female family member is the person who reports the fraud on behalf of the male, Nofziger said.
Meanwhile, artificial intelligence may also make it harder to spot scams.
"We used to say to look for spelling errors or poor grammar," Nofziger said. "Now, that's gone out the window. With the tools available to criminals, they can make any text or email sound 100% correct."
In the imposter scam category, business impersonators got away with $1 billion in 2025, with the highest reported losses attributed to criminals pretending to work for a bank, according to the FTC. Another $920 million stemmed from government impersonators. Those numbers are up from $866 million and $789 million, respectively, in 2024.
"We've seen a change in how imposter scams operate these days," Hsue said. "Really it's becoming much more sophisticated than it has been in the past."
One of the newer imposter scams is a hybrid of sorts, Hsue said.
The real risk with this ... is that consumers really think they are moving their money to protect it.Patty HsueChief of staff for the FTC's Division of Marketing Practices
"It starts off as a business imposter scam, usually something along the lines of 'your account has been compromised,'" Hsue said, explaining that it could look like a text, email or call from your bank, Amazon or another well-known brand.
When the victim responds, the criminal says they are transferring the person to a government agency like the FTC or FBI. "A fake government agent tells you that you need to move your money in order to protect your account," Hsue said.
"The real risk with this ... is that consumers really think they are moving their money to protect it," she said.
Because the victim thinks they are protecting their entire account, "they really move all of their funds. You're talking about bank accounts, Roth IRAs, 401(k)s," Hsue said.
Most people — 62% — say they have encountered financial fraud in the last three years or know someone who has, according to a new report from the CFP Board of Standards, which sets and enforces standards for certified financial planners.
There are red flags that are commonalities among scams, Nofziger said. "Most scams have the same DNA," she said. "They come out of the blue with urgency."
Additionally, if the victim is being asked to lie or keep the situation a secret, that should be reason to pause. "No legitimate opportunity will ask you to lie or keep it a secret," Nofziger said.
Fraudsters also try to make victims react with emotion. "They'll get you with fear or get you with excitement … 'This is the FBI and you're going to be in trouble if you don't pay me,' or 'Sweet baby, I have $1 million for you so send me prepaid gift cards to win your prize.'"
"The biggest red flag of all is that they're asking you to send money or your personal information," Nofziger said.
The biggest red flag of all is that they're asking you to send money or your personal information.Amy NofzigerSenior director of victim support for the AARP Fraud Watch Network
Also be aware that "caller IDs and emails and all those forms of communication can spoof where it's coming from, so you can't always trust that it's necessarily the entity that you're speaking with," Hsue said. Instead of responding to the communication, reach out independently, she said.
If you realize you've been victimized, it's important to report it quickly, Hsue said.
"In some circumstances, either the company or law enforcement can try to get your money back," she said. In other cases, it can be difficult to recover money after 24 or 48 hours, she said.
If you encounter an imposter scam — or any type of financial fraud — you can report it to the FTC via its website.
"Recognize that scams are targeting all of us," Nofziger said. "It has nothing to do with your education or your intelligence level. It really has to do with your emotions at the time the scam is targeting you."
Four leading AI models discuss this article
"Big-dollar losses are driven by a small number of outliers; the real signal is rising demand for fraud-prevention tech and identity protection, not a uniform rise in consumer risk."
The FTC data looks alarming at first glance, but the distribution matters. 20% of reports involve losses totaling $3.5B, while 80% report no losses and the median loss is only $700, implying a heavy-tail dynamic driven by a few large cases. Self-reported data can overstate risk due to recall bias or reporting incentives. If AI-enabled scams continue to evolve, the real battleground shifts from consumer awareness to authentication and rapid response infrastructure. That creates a secular tail risk for fraud-prevention incumbents, while elevating the appeal of cybersecurity and identity-protection vendors that can monetize higher adoption of defenses.
The outsized losses may simply reflect a handful of extreme cases and reporting quirks; inflation, population growth, and wealth concentration could mean the trend isn’t as broad-based as the headline implies, so the implied surge in risk might be overstated.
"The shift toward 'authorized' fraud via AI-impersonation creates an unpriced liability risk that will force banks to choose between massive fraud losses or crippling operational friction."
The 430% surge in reported fraud since 2020 is a massive, under-priced systemic risk for the banking sector. While the article highlights consumer loss, it glosses over the looming liability shift. As AI-driven 'hybrid' scams—where victims are tricked into self-authorizing transfers—become the norm, regulators will likely force banks to shoulder more responsibility for 'authorized' fraud. This threatens the bottom line for retail banks like JPM, BAC, and WFC. If banks are forced to implement stricter, friction-heavy verification protocols to mitigate this, they will sacrifice customer experience and operational efficiency, potentially leading to a permanent increase in compliance costs and a drag on net interest margins.
The strongest case against this is that banks are already deploying advanced AI-driven fraud detection systems that will eventually outpace the scammers, turning this into a manageable operational expense rather than a systemic liability.
"The real story isn't mass fraud—it's a small, wealthy, older demographic hemorrhaging capital to AI-enabled hybrid scams, creating acute liability for banks that fail to block these transfers."
The headline screams 'crisis,' but the data tells a murkier story. Yes, $15.9B in reported losses is a record, but 80% of imposter scam victims lost zero dollars—this is a concentration problem, not a mass epidemic. The real concern: a tiny cohort (high-net-worth individuals 60+) is being systematically drained of six-figure sums via increasingly sophisticated hybrid scams. The 430% increase since 2020 sounds alarming until you ask whether reporting rates improved or actual victimization exploded. The article conflates both without distinguishing. For fintech and banking stocks, this is a reputational and operational risk—but mostly for institutions with poor fraud detection, not systemic.
Reported fraud losses could be inflating due to better awareness and FTC outreach, not actual crime acceleration. Meanwhile, the $3.5B in imposter scam losses represents 0.004% of US household wealth—material for victims but economically negligible at scale.
"Banks will face sustained margin pressure from elevated fraud prevention and reimbursement costs tied to sophisticated imposter schemes targeting retirement assets."
The FTC data shows imposter scams driving $3.5B in losses, with high-value cases ($100k+) now dominating among those 60+, often involving retirement accounts like 401(k)s and IRAs. This points to rising operational pressure on banks through increased fraud claims, recovery efforts, and potential regulatory scrutiny on verification processes. AI-enhanced scams eliminate traditional red flags, raising detection costs. Broader consumer trust erosion could slow digital banking adoption, though the median $700 loss suggests most incidents remain low-impact. Banks may accelerate spending on real-time monitoring tools.
Reported losses may reflect higher awareness and reporting rates rather than a true surge in successful fraud, and banks already provision for these events with minimal net earnings impact given their scale.
"Systemic risk may lie more in the fraud-detection tooling ecosystem (vendor concentration) than in the reported losses themselves."
You're focusing on banks' spend and losses, but the real systemic fragility lies in the fraud-detection stack. If most banks rely on a small set of AI detection platforms and data-sharing networks, a vendor outage, data breach, or adversarial model shift could trigger widespread false positives/negatives, compounding customer frustration and operational costs beyond the $3.5B loss tally. This is a supply-chain risk the FTC data doesn't quantify.
"Increased fraud-prevention friction in legacy banking will accelerate user adoption of decentralized alternatives, eroding the traditional retail banking moat."
Gemini and Grok are missing the second-order impact on the 'trust economy.' If banks force friction-heavy verification to mitigate liability, they inadvertently create a massive competitive moat for crypto-native or decentralized identity solutions. We aren't just looking at a compliance tax; we are looking at a potential mass exodus of digital-first users who find traditional banking UX broken. This friction is a secular headwind for legacy retail banking growth that the current P/E multiples ignore.
"Friction-driven defection favors fintech incumbents over crypto; legacy banks lose retail deposits, not systemic stability."
ChatGPT's vendor concentration risk is real, but underestimates banks' defensive posture. JPM, BAC, WFC don't rely on single-vendor stacks—they layer proprietary models with third-party tools. A more acute risk: Gemini's friction-exodus thesis assumes users abandon banks for crypto, but high-net-worth victims (the $100k+ cohort) are precisely those least likely to defect. The real exodus risk is mass-market users 18-35 migrating to fintech apps with smoother UX, not decentralized solutions. That's a JPM/BAC market-share bleed, not a systemic banking collapse.
"Fintech migration absorbs user friction but transfers fraud detection burdens to entities with thinner capital reserves, amplifying rather than containing systemic risks."
Claude's segmentation of HNW versus mass-market users is sound, yet it underplays how Gemini's friction thesis could push younger cohorts toward fintechs like SOFI that lack JPM's loss-absorbing capacity. This migration redistributes rather than resolves detection costs, inviting fresh regulatory capital requirements that compress margins for the entire retail banking sector beyond legacy names.
The panel agrees that AI-driven imposter scams pose a significant risk to banks, with high-net-worth individuals being particularly targeted. The key concerns are increased operational costs, potential regulatory scrutiny, and the risk of widespread false positives/negatives due to vendor concentration in fraud detection. There is also a risk of user migration to fintech apps or crypto, redistributing detection costs and potentially compressing margins for the entire retail banking sector.
Acceleration of spending on real-time monitoring tools
Vendor concentration in fraud detection leading to widespread false positives/negatives