I built an AI scam detection platform after my mom got deepfaked

The first time I got scammed, I was a college student buying a laptop on Craigslist. The seller took my money and disappeared. Expensive lesson, but survivable.

The second time hit different. My mom got a call from someone who sounded exactly like her cousin (at least according to her). Same voice, same mannerisms. It was AI voice cloning. That call was the tipping point. 

I spent 25 years building enterprise software (fraud detection at Tagged/Myspace). I knew the data existed to fight this. The FTC and FCC has millions of consumer complaints. Malware threat feeds track malicious domains. But none of it was accessible to regular people.

So I built ScamVerify. You enter a phone number, website, text message, or email, and it pulls from all those government and threat intelligence sources, runs AI analysis, and gives you a plain-English verdict with a 0-100 risk score.

Some things I found in the data that surprised me:

• 345,670 FTC complaints are for debt reduction robocalls alone, with an 89% robocall rate

• We identified 15 scam rings operating across multiple phone numbers 

• AI voice cloning needs just 3 seconds of audio from a social media video to clone someone's voice 

• 154,716 impersonation complaints in the FTC database, and the ones using live agents are the most likely to deploy AI cloning        

The consumer product is free at scamverify.ai. We also launched a B2B API for fintech and fraud teams that need this data programmatically. Would love feedback from this community.