What Happened
Malwarebytes published its 2026 report on AI and scams, which included survey data showing that nearly nine in ten adults (85%) say they can no longer distinguish real images from AI-generated ones. That's a massive jump from just 66% in 2025 — a 19 percentage point shift in a single year.
This is not a theoretical concern. It's the foundation of every major scam tactic deployed in 2026: deepfake videos, synthetic celebrity endorsements, fake news photos, AI voice cloning, and photorealistic lies spread at the speed of social media.
Why This Matters
For the past seventy years, visual evidence was considered the gold standard of proof. You could lie in writing, but a photograph was hard to fake. Video was even harder. A video of something happening was essentially proof that it happened.
That's gone. AI image generation is now good enough that photorealism is no longer proof of authenticity. For the average person, there is no longer a reliable way to tell what's real and what's generated.
The security industry's traditional advice — "trust your eyes" — is now obsolete.
The Scam Infrastructure
This capability is now being weaponized at scale:
Deepfake Videos: AI video of a celebrity promoting a crypto scheme, a politician making a damaging statement, a friend asking for emergency money — all synthetic, all convincing enough to fool people who are in a hurry.
Fake News Images: Photos of disasters, political events, or celebrity scandals that never happened. Shared millions of times before anyone verifies the source.
AI Voice Cloning: Scammers now record short samples of people's voices and use AI to generate full sentences and conversations. Your mother calling to say she's in jail? Might be AI.
Synthetic Identities: AI-generated photos of people that don't exist, used to build trust in scam relationships over weeks or months. By the time the person realizes it's fake, they've already transferred money.
The Real Stupid Shit
The real stupid shit is that we built all this technology and shipped it to the public without any infrastructure for verification or authentication. No digital watermarks that survive compression. No tamper-proof certificates of authenticity. No easy way for a regular person to verify that what they're looking at is real.
Instead, we have a world where 85% of adults admit they can't tell the difference between reality and a computer hallucination. Scammers have that data point. They're using it.
What's the Solution?
Short answer: nobody knows yet. The technology moved faster than the defenses. Malwarebytes, Google, the FBI, and every major tech company are scrambling to deploy detection tools and verification systems. But if people can't trust their own eyes, how are they supposed to trust a third-party verification tool?
This isn't a problem that gets solved in 2026. This is the underlying condition for every scam, hoax, and piece of misinformation that's coming.
Sources
Help Net Security: 9 out of 10 people can no longer distinguish real from AI-generated content
Malwarebytes: 2026 Face Value Report on AI Scams
Malwarebytes: Americans lost nearly $900 million to AI-powered scams, FBI says