New AI tricks make crypto scams harder to spot

A Trap Disguised as a Routine Call

A tech founder trusted what looked like a legitimate Microsoft Teams call—complete with familiar faces and voices. Two other participants joined, creating the illusion of a normal meeting. Then, the video froze. A prompt appeared, claiming his software needed an urgent update. He followed the instructions, shut his laptop, and only later realized he had been tricked.

This wasn’t just a phishing email or a suspicious link—it was a sophisticated deepfake scam that leveraged AI to impersonate people he knew. The attackers mimicked voices and faces so convincingly that trust alone became the vector of attack.

The Evolution of Social Engineering: From Text to Video

Social engineers have long relied on manipulation, but video calls have democratized deception. Where scams once required weeks of text-based grooming, today’s AI can replicate a person’s voice and appearance in minutes.

• Microsoft’s Warning (2026): Fake Teams and Zoom files disguising malicious updates.
• Google’s Discovery: Spoofed Zoom meetings paired with AI-generated executive videos.
• The New Normal: Deepfakes and synthetic media are making scams faster, cheaper, and harder to detect.

The Cardano Foundation later confirmed their Telegram account was hacked—but the damage was already done. The scammer continued the ruse for days, even asking to reschedule meetings, proving how these attacks can unfold over time.

Why Crypto? Why Now?

Cryptocurrency remains a prime target because transactions are irreversible and fast. Victims trust rapid video calls, believing they’re communicating with colleagues or partners. According to Chainalysis, crypto fraud cost victims $17 billion in 2025—with AI-powered scams generating 4.5 times more revenue than traditional methods.

But the threat doesn’t end with stolen funds. Attackers use compromised data to launch new waves of fraud, turning one breach into a chain reaction of attacks.

The Counterattack: Can Defenses Keep Up?

Some companies are deploying verification tools to fight back:

• Zoom’s “Verified Human” badge (a step toward authenticating real users).
• Pre-agreed passphrases, secondary verification channels.

Yet, despite these innovations, Gartner estimates half of businesses won’t adopt such defenses until 2027. Meanwhile, AI is making scams faster, cheaper, and more scalable—leaving victims with fewer ways to fight back.

The Hard Truth: Trust No One, Verify Everything

One convincing deepfake could mean the difference between safety and disaster. Never rely on video or voice alone. Always use secondary verification methods—call back on a trusted number, use cryptographic signatures, or implement multi-factor authentication.

In a world where seeing isn’t believing, the best defense is skepticism.