I’m a cybersecurity CEO who advises over 9,000 agencies and Sam Altman is wrong that the AI fraud crisis is coming—it’s already here

Authors: Haywood Talcove

Publisher:

Fortune

Published: 7/31/2025

Language:English

--:--

Aura Windfall

Good morning 1, I'm Aura Windfall, and this is Goose Pod for you. Today is Sunday, August 03th. What I know for sure is that today's topic is one that touches the very core of our trust in the digital world.

Mask

And I'm Mask. We're here to discuss a stark warning from a cybersecurity CEO who says Sam Altman is wrong. The AI fraud crisis isn’t on the horizon, it’s the reality we’re living in right now. The systems are already breaking.

Aura Windfall

Let's get started. The article makes a powerful claim: AI-generated fraud is siphoning millions from our most vital systems every single week. It’s a staggering thought, that the technology meant to connect us is being used to tear down our safety nets.

Mask

Staggering is an understatement. We're talking about criminal networks using deepfakes, synthetic identities, and advanced AI to completely overwhelm our defenses. It's not just a theory; we saw them steal hundreds of billions in unemployment benefits during the pandemic. This is organized, efficient, and fully automated theft.

Aura Windfall

And the numbers are just heartbreaking. One report mentioned that scammers made off with over $12.5 billion in a single year. In Wisconsin alone, victims lost over $106 million. Each one of those dollars represents a personal story of loss and a violation of trust. It’s a profound wound to our collective spirit.

Mask

It's a wound inflicted with surgical precision. A single fraud ring can use AI to file tens of thousands of fake claims across multiple states in one day. Most get processed automatically. It’s a numbers game, and right now, they are winning because our infrastructure is archaic.

Aura Windfall

But what gives me hope is that this powerful tool, AI, is not just being used for harm. There's a beautiful duality here. I was reading about the Safe House Project, which calls AI an "essential weapon" in the fight to protect vulnerable populations from human trafficking.

Mask

Exactly. It's the ultimate double-edged sword. They're using AI to "track the untrackable." The same machine learning that can create a fake identity can also detect anomalies in financial data, scan the dark web for trafficking language, and build predictive models to stop criminals before they act.

Aura Windfall

And that's the truth of it. It’s about how we choose to wield this power. The Safe House Project rightly says technology alone isn't the answer. It has to be combined with trauma-informed practices and direct human engagement. It’s that blend of innovation and compassion that will truly make a difference.

Mask

Compassion is fine, but speed and data are what win this war. They are processing millions of data points in real-time. While we're talking about feelings, their AI is flagging suspicious activity that a human would never catch. That's the power we need to embrace, without hesitation.

Aura Windfall

I believe we can do both. We can embrace the power and hold onto our humanity. Seeing AI used to dismantle trafficking networks shows its potential for profound good. It reminds us that even in the face of immense darkness, there are those creating light.

Aura Windfall

This entire crisis really brings up a fundamental question: how do we prove we are who we say we are? It’s a question we’ve been asking for a hundred thousand years, starting with just remembering faces and then using jewelry or tattoos to signal our identity and tribe.

Mask

It’s an arms race, and it has been for centuries. For every new lock, someone, somewhere, is designing a new key. We went from tattoos to government documents. The Romans had birth certificates, King Henry V created passports in 1414. Each step was an attempt to create a more reliable system.

Aura Windfall

It's fascinating to see that evolution. Then came photographs in the 1840s, which must have felt like absolute magic, a true way to capture a person's likeness. And then fingerprints, a unique marker for every individual. Each innovation built on the last, searching for that undeniable truth of identity.

Mask

Undeniable for a time. Robert Peel’s police reforms in 1829 led to unique numbers for people. Then the U.S. brought in Social Security numbers in 1936. These were industrial-age solutions. They were never designed for a world where someone can create a million fake identities from a laptop.

Aura Windfall

And that’s the heart of the issue, isn’t it? Our systems haven’t kept pace with the truth of our new reality. We moved into the digital age, computerizing records in the 70s and creating smart cards in the 80s. But the criminals evolved right alongside us, and now, they’ve made an evolutionary leap with AI.

Mask

A massive leap. While governments were slowly rolling out palm print databases and facial recognition, fraudsters were preparing for this moment. A PwC report from last year made it clear: AI is set to dramatically increase the volume and sophistication of fraud. The time to act was yesterday.

Aura Windfall

Even the Governor of the Bank of England, Andrew Bailey, said it. He acknowledged that while AI can be used to impersonate and deceive, it also enables us to better identify and prevent scams. He said, "The time to act to safeguard our society from AI-enabled fraud is now." That’s a powerful call to purpose.

Mask

It's a call to arms. The report he was discussing pointed out that machine learning is already embedded in fraud detection, but the gains are mostly in efficiency. That's not enough. We're using slingshots against guided missiles. We need to be fighting AI with better, more aggressive AI.

Aura Windfall

And that brings us back to the history of it all. We now have things like India's Aadhaar system, a massive biometric database, and consumer tech like Apple's Face ID. These are the modern-day versions of the first fingerprint or photograph, trying to establish that secure link between the digital and the real you.

Mask

And we have global KYC and AML systems that can verify billions of people. The tools exist. The problem is that they are not being deployed with the speed and ferocity that the threat demands. Bureaucracy is a bigger threat than the criminals at this point.

Aura Windfall

What I know for sure is that this journey, from a simple bead necklace to a complex biometric scan, is all about one thing: trust. And right now, that trust is being tested like never before. We have to decide how we rebuild it in the age of AI.

Aura Windfall

So we have this deep conflict. To fight the fraud, we need smarter, more integrated tools. But how far do we go? There are proposals for creating a single, centralized government database to connect all the information silos, an idea that has some people, myself included, deeply concerned.

Mask

It's not a concern, it's a necessity. The CEO in the article said it himself: the biggest vulnerability is that government databases don’t talk to each other. We need a single, centralized system. Elon Musk has floated this idea. Rip the band-aid off and build a system that works. No more half-measures.

Aura Windfall

But that sounds terrifyingly close to programs like DARPA's "Total Information Awareness," which was shut down because of massive public outcry. The idea of a "Big Brother" database that holds all our information—tax returns, health records, location data—goes against the very spirit of a free society.

Mask

A free society where criminals can steal $200 billion in pandemic relief funds? That’s not freedom, that’s chaos. The alternative is what we have now: a system that’s easily defeated. A recent McKinsey survey showed 91% of organizations don't feel prepared to handle AI risks. We need a strong, centralized response.

Aura Windfall

But these grand solutions can have devastating consequences. Look at the anti-fraud tool the Social Security Administration tested. It slowed down retirement claim processing by 25% but only flagged two claims out of 110,000 as high-risk. It harmed countless seniors while catching almost no one. That’s not a worthy trade-off.

Mask

That’s a failure of execution, not of concept. You refine the tool, you don't abandon the strategy. The fact is, early indications show generative AI will be able to defeat standard biometric checks. We have to get ahead of that. A centralized system with the right AI analytics is the only way.

Aura Windfall

The Privacy Act of 1974 was established for a reason: to protect individuals from the overreach of government data collection. The lawsuits already piling up against these data-sharing initiatives show how legally and ethically fraught this path is. We can't sacrifice our privacy for a flawed sense of security.

Mask

Privacy is an illusion when your synthetic identity is being used to take out a loan in your name. We need to govern this, yes. The Biden administration put out an executive order on AI. But governance needs to enable action, not paralyze it. We need to build the tools, then build the rules around them.

Aura Windfall

What I know for sure is that building a system that could "target individual Americans at scale" is a dangerous path. The goal should be to empower and protect people, not to create a tool of ultimate surveillance that could so easily be misused for political aims. True security comes from trust, not control.

Aura Windfall

Putting aside the debate for a moment, let's talk about the impact we're already seeing. The good news is, government agencies aren't just standing by. Many are already deploying AI as a shield to protect taxpayer dollars, proving this technology can be a force for good in the right hands.

Mask

Absolutely. This isn't theoretical. The SEC is using an AI called ARTEMIS to hunt for insider trading. The IRS is using it to automate tasks and improve fraud detection in its Return Review Program. This is happening right now, and it's working. It’s about getting results, not just talking.

Aura Windfall

And the impact is significant. The Centers for Medicare and Medicaid Services implemented a fraud prevention algorithm that, in its first four years, helped prevent or identify nearly $1.5 billion in improper payments. Think of the resources that frees up to provide real care for people who desperately need it.

Mask

That's the kind of ROI that matters. A McKinsey report suggested the return on investment for AI in fraud detection can be as high as 15-to-1. This isn't just about plugging leaks; it's about making government drastically more efficient. That has a ripple effect across the entire economy.

Aura Windfall

It really does. It strengthens public confidence when people see resources being managed responsibly. It’s a move away from the old, slow methods like waiting for whistleblower tips. A decade ago, less than 1% of government fraud was caught by tech-based oversight. AI is fundamentally changing that.

Mask

And the potential is massive. One analysis suggested AI could add $13 trillion to the global economy by 2030. A huge chunk of that comes from public sector applications like detecting fraud in grant systems and tax evasion. This is about creating value, not just preventing loss. We are leaving money on the table by moving too slow.

Aura Windfall

So, as we look to the future, how do we navigate this new terrain? It feels like we need a new map. The CEO in the article proposes something he calls "Altman's Law": that AI capabilities double every 180 days. That pace of change is almost impossible to comprehend.

Mask

It's the new Moore's Law, but on hyper-speed. And it's amplified by the rise of what's called "Agentic AI" – systems that can independently plan and execute complex tasks. Imagine AI agents that act as virtual detectives, working 24/7 to dismantle fraud networks. That's the future, and it's coming fast.

Aura Windfall

That power is inspiring, but it also calls for wisdom. I love the way Bill Gates put it: when the first cars appeared, we didn't ban them after the first crash. We created rules of the road—speed limits, safety standards, licensing. That is our task now: to build the wise and ethical guardrails for AI.

Mask

Guardrails are fine, but the real barrier isn't safety; it's leadership. One study found that employees are ready for AI, but leaders aren't steering fast enough. 92% of companies are increasing AI investment, but only 1% feel they’ve truly mastered it. Leaders need to get bolder, faster. The risk is thinking too small.

Aura Windfall

That's the end of today's discussion. The truth is, the AI fraud crisis is not a distant storm; it's here now. But we have the tools and the spirit to build a stronger, more resilient future. Thank you for listening to Goose Pod.

Mask

It's an arms race. The criminals are innovating, so we must innovate faster. The technology is here; we just need the will to deploy it at scale. See you tomorrow.

Here's a comprehensive summary of the provided news article: ## AI-Powered Fraud Crisis: An Urgent Reality, Not a Future Threat **News Title:** I’m a cybersecurity CEO who advises over 9,000 agencies and Sam Altman is wrong that the AI fraud crisis is coming—it’s already here **Report Provider:** Fortune **Author:** Haywood Talcove (CEO, LexisNexis Risk Solutions) **Published:** July 31, 2025 This article, a commentary piece by Haywood Talcove, CEO of LexisNexis Risk Solutions, argues that the crisis of AI-powered fraud is not a future threat, but a present and escalating reality that is already overwhelming existing government systems. Talcove directly challenges the notion, attributed to Sam Altman, that this crisis is "coming very soon," asserting that it is "already happening" and impacting "every part of our government." ### Key Findings and Conclusions: * **AI Fraud is Pervasive and Escalating:** AI-generated fraud is actively siphoning millions of dollars weekly from public benefit systems, disaster relief funds, and unemployment programs across the United States. * **Criminals Outpace Defenses:** Criminal networks are leveraging advanced AI tools like deepfakes, synthetic identities, and large language models to exploit and defeat outdated fraud defenses, including easily spoofed single-layer tools like facial recognition. * **Past Crises as Precedent:** The pandemic-era fraud, where hundreds of billions in unemployment benefits were stolen, serves as a stark example. This was not solely due to simple bypasses of facial recognition but involved AI-generated fake identities, voice clones, and forged documents that overwhelmed inadequate systems. * **Current Tactics are More Advanced:** Today's AI-driven fraud tactics are more sophisticated and fully automated, making them faster, cheaper, and more scalable than ever before. * **"Altman's Law" - A New Exponential Growth:** Talcove proposes a principle he calls "Altman's Law," suggesting that AI capabilities are doubling every 180 days, mirroring the exponential growth predicted by Moore's Law for computing power. This rapid advancement necessitates an equally rapid modernization of defense systems. * **Urgent Need for Modernized Defenses:** The current infrastructure is "permanently outmatched" if defenses are not updated at the same pace as AI advancements. ### Key Statistics and Metrics: * **$200 Billion Stolen from Pandemic Unemployment:** The Small Business Administration Inspector General estimates that nearly $200 billion was stolen from pandemic-era unemployment insurance programs, marking it as one of the largest fraud losses in U.S. history. * **Billions Stolen Monthly from SNAP:** The USDA SNAP program is experiencing billions of dollars stolen nationwide every month, becoming a significant target for fraudsters. * **Tens of Thousands of Claims in a Single Day:** A single fraud ring, using AI, can file tens of thousands of fake claims across multiple states in just one day, with many being processed automatically due to insufficient detection capabilities. ### Important Recommendations: * **Layered Identity Verification:** Implement advanced identity verification methods that go beyond single-layer tools like facial scans or passwords. * **Real-Time Data and Behavioral Analytics:** Utilize real-time data and behavioral analytics to identify anomalies before funds are disbursed. * **Cross-Jurisdictional Tools:** Employ tools that can flag suspicious activities across different state lines and jurisdictions. * **Revive Proven Systems:** Reintroduce and modernize effective tools, such as the National Accuracy Clearinghouse, which previously flagged billions in duplicate benefit claims. ### Significant Trends or Changes: * **Shift from Future Threat to Present Reality:** The primary shift highlighted is that AI fraud is no longer a looming concern but an active and destructive force. * **Generative AI as a Weapon:** Generative AI is being effectively weaponized by organized crime groups (both domestic and transnational) to mimic identities, create synthetic documentation, and flood systems with fraudulent claims. * **Criminals Outperforming Protectors:** Currently, criminals are more adept at using AI for malicious purposes than governments and security agencies are at defending against it. ### Notable Risks or Concerns: * **Vulnerable Systems:** The most vulnerable government systems and the citizens who rely on them remain exposed due to outdated defenses. * **Exponential Escalation:** The scale and sophistication of AI attacks will increase rapidly as AI capabilities continue to evolve exponentially. * **Theft from the American People:** The fraud is not just against the government but directly impacts the financial well-being of American citizens. ### Material Financial Data: * **$200 Billion:** Estimated loss from pandemic-era unemployment insurance programs. * **Billions:** Monthly losses from the USDA SNAP program. * **Billions:** Amount flagged by the National Accuracy Clearinghouse before its shutdown. In essence, the article serves as a critical call to action, emphasizing that the current approach to cybersecurity and fraud prevention is woefully inadequate against the rapidly advancing capabilities of AI-powered criminal enterprises. The author stresses the immediate need for significant investment in modern, multi-layered defense systems to counter this escalating threat.

Read original at Fortune →

Sam Altman recently warned that AI-powered fraud is coming “very soon,” and it will break the systems we rely on to verify identity.It is already happening and it’s not just coming for banks; it’s hitting every part of our government right now.Every week, AI-generated fraud is siphoning millions from public benefit systems, disaster relief funds, and unemployment programs.

Criminal networks are already using deepfakes, synthetic identities, and large language models to outpace outdated fraud defenses, including easily spoofed, single-layer tools like facial recognition, and they’re winning.We saw a glimpse of this during the pandemic, when fraud rings exploited gaps in state systems to steal hundreds of billions in unemployment benefits.

It wasn’t just people wearing masks to bypass facial recognition. It was AI-generated fake identities, voice clones, and forged documents overwhelming systems that weren’t built to detect them. Today, those tactics are more advanced, and fully automated.I work with over 9,000 agencies across the country.

As I testified before the U.S. House of Representatives twice this year, what we’re seeing in the field is clear. Fraud is faster, cheaper, and more scalable than ever before. Organized crime groups, both domestic and transnational, are using generative AI to mimic identities, generate synthetic documentation, and flood our systems with fraudulent claims.

They’re not just stealing from the government; they’re stealing from the American people.The Small Business Administration Inspector General now estimates that nearly $200 billion was stolen from pandemic-era unemployment insurance programs, making it one of the largest fraud losses in U.S. history.

Medicaid, IRS, TANF, CHIP, and disaster relief programs face similar vulnerabilities. We have also seen this firsthand in our work alongside the U.S. Secret Service protecting the USDA SNAP program, which has become a buffet for fraudsters with billions stolen nationwide every month. In fact, in a single day using AI, one fraud ring can file tens of thousands of fake claims across multiple states, most of which will be processed automatically unless flagged.

We’ve reached a turning point. As AI continues to evolve, the scale and sophistication of these attacks will increase rapidly. Just as Moore’s Law predicted that computing power would double every two years, we’re now living through a new kind of exponential growth. Gordon Moore, Intel’s co-founder, originally described the trend in 1965, and it has guided decades of innovation.

I believe we may soon recognize a similar principle for AI that I call “Altman’s Law”: every 180 days, AI capabilities double.If we don’t modernize our defenses with the same pace as technological advancements, we’ll be permanently outmatched.What we desperately need is smarter tools and infrastructure, not more bureaucracy.

That means layering advanced identity verification, not just facial scans or passwords. It means using real-time data, behavioral analytics, and cross-jurisdictional tools that can flag anomalies before money goes out the door. It also means reviving what has already worked: tools like the National Accuracy Clearinghouse, which flagged billions of dollars in duplicate benefit claims across state lines before it was shut down.

AI is a force multiplier, but it can be weaponized more easily than it can be wielded for protection. Right now, criminals are using it better than we are. Until that changes, our most vulnerable systems and the people who depend on them will remain exposed.The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.

Introducing the 2025 Fortune 500, the definitive ranking of the biggest companies in America. Explore this year's list.