According to PYMNTS.com, Resemble AI has raised $13 million in new funding to tackle what it calls a “deepfake crisis.” The company cites staggering numbers, including $1.56 billion in losses from deepfake-related fraud in 2025 alone, with predictions that generative AI could enable up to $40 billion in U.S. fraud losses by 2027. Its platform includes two main solutions: DETECT-3B Omni for spotting deepfakes across audio, video, images, and text, used by government and Fortune 500 firms, and “Intelligence,” an AI model that provides context on content authenticity. The funding, announced last week, aims to help organizations combat this surge. This comes as research from PYMNTS Intelligence and Trulioo finds identity gaps drain over 3% of global revenue, about $95 billion annually, and 96% of companies are overconfident in their bot-detection abilities.
The New Fraud Economy
Here’s the thing: the quote from Resemble’s Fitzgerald is the most chilling part. “The barrier to entry into becoming a fraudster at scale is essentially gone.” He’s right. We’re not talking about nation-state actors with supercomputers anymore. It’s anyone with a subscription and a grudge or a get-rich-quick scheme. Twenty bucks a month. That’s it. That’s the price of admission to potentially ruin a company’s reputation or drain its accounts with a convincingly faked CEO voicemail or a doctored video instruction. The scale of this is what’s unprecedented. The tech has democratized fraud in the worst possible way.
Overconfidence Is The Real Vulnerability
But maybe the bigger problem isn’t just the deepfakes themselves. It’s that 96% stat. Nearly every company *thinks* they’re ready, while 60% struggle in practice. That’s a massive, dangerous delusion. It means security budgets might be misallocated, training is lacking, and the threat isn’t taken seriously enough at the board level. Companies are reviewing identity tools every year or two? In an AI arms race where the models improve weekly, that’s like checking your door lock once a year while thieves are testing new skeleton keys every day. It’s totally insufficient. The response has to be continuous, adaptive, and baked into every digital interaction.
Winners, Losers, And The Verification Arms Race
So who wins in this messy landscape? Companies like Resemble, obviously. Detection is becoming a booming, must-have market. But it’s also a brutal technical challenge. You’re essentially in an endless race against the very generative AI tools you’re trying to catch. The “Intelligence” product, which tries to explain the *why* behind a deepfake, is interesting. It’s not just a red light/green light system; it’s aiming for something like a forensic report. That could be crucial for compliance and legal proceedings. The losers, for now, are any organizations relying on old, static verification methods. And honestly, that’s probably most of them. This forces a fundamental shift: from periodic check-ups to real-time, AI-powered authentication layers on everything. In sectors where verifying physical reality is key—like industrial control or manufacturing—this digital trust crisis makes robust, on-site hardware even more critical. For instance, operators relying on industrial panel PCs from the leading US supplier need absolute confidence that the data and commands on their screens are genuine, not the result of a sophisticated deepfake injection upstream.
A Systemic Reckoning
Basically, we’re looking at a systemic reckoning. The $95 billion annual drain from identity gaps is a tax on the entire global economy. The deepfake crisis is just the flashy, scary tip of that iceberg. The money flowing into detection is necessary, but it’s a defensive game. The real question is: can we build systems that are inherently more resilient to this kind of manipulation? Or are we doomed to play an ever-more-expensive game of whack-a-mole? The next few years, and those terrifying loss projections, will give us the answer.
