Updated on: Sept 01, 2025 06:03 pm IST
AI impersonation scams are surging in 2025, with fraudsters cloning voices and faces to trick victims. Here’s how you take to stay safe online.
Voice-clone scams are no longer sci-fi. In 2025, impersonation fraud using AI is exploding, thanks to realistic deepfake video, cloned speech, and highly tailored messages.
According to Moonlock, such scams surged by 148% this year. That’s not hype, it’s a clear signal that deception is evolving fast.
How the impersonation game has changed
- Voice cloning is now alarmingly accurate. With just a few seconds of audio from social media, podcasts, or voicemails, AI can replicate a voice so closely that even trained ears may struggle to tell the difference.
- Deepfake video calls are becoming a real threat. Fake CEOs, mentors, or high-level managers appear in pixel-perfect video meetings and order wire transfers without raising suspicion.
- Hyperpersonalized phishing is the new normal. AI tools gather information from social profiles to craft messages packed with personal context. That familiarity lowers defences, and quickens the hand that sends money
Why it’s a problem
Criminals are using increasingly accessible AI to create believable fraud fast, and at scale. The U.S. alone could lose up to $40 billion annually by 2027 to AI-enabled scams. One standout case: an employee in Hong Kong was fooled in a video call by a deepfake CFO and authorized a $25 million transfer
How to stay safe
Action | Benefit |
| Pause and verify | Always call back using a verified number before acting. |
| Spot over-detailing | Messages packed with unnecessary personal context are suspects.(turn0search3) |
| Stick to trusted platforms | Avoid clicking unknown links, even if they come from “trusted” voices. |
| Use multi-factor authentication (MFA) | Adds a safety net if impersonation fails. |
AI impersonation scams are climbing, and they’re targeting trust itself. Seeing a known face or hearing a familiar voice is no guarantee anymore. The only reliable defence is scepticism, verification, and controls that force scammers to slow down.
