Deepfake Scams: Every Major Type, Real Cases, and How to Protect Yourself

---

In February 2024, an employee at engineering firm Arup wired $25.6 million to criminals after a video meeting where every colleague on screen, including the CFO, was a deepfake. That single case explains why deepfake scams have become the fastest-evolving fraud category in the world.

This page is the complete map. Every major type of deepfake scam, a documented real case for each, and a link to our in-depth guide on every one. Bookmark it, share it with your team, and use the verification workflow before you ever act on an urgent video, voice, or image.

What is a deepfake scam? Deepfake scams use AI-generated video, audio, or images to impersonate real people for fraud. The major types are executive fraud, voice cloning calls, investment scams, romance scams, phishing, and identity fraud. Spot them by verifying through a second channel and running media through a deepfake detector.

What Counts as a Deepfake Scam?

A deepfake scam is any fraud that uses synthetic media, AI-generated faces, voices, or video, to impersonate a real person or fabricate a believable one. Classic impersonation needed acting skill. Deepfakes need only source material: a LinkedIn headshot, a conference talk, three seconds of voicemail.

The losses are no longer hypothetical. The FBI's 2025 IC3 report introduced AI-assisted cybercrime as a formal category for the first time, logging over 22,000 complaints and nearly $900 million in losses in a single year. Total reported cybercrime losses hit $20.9 billion in 2025, up 26 percent from 2024.

And the trend line points up. Deloitte's Center for Financial Services projects that generative AI could push US fraud losses to $40 billion by 2027, up from $12.3 billion in 2023. For the full sourced dataset, see our deepfake statistics page.

The 8 Types of Deepfake Scams (With Real Cases)

1. Executive and CEO fraud deepfakes (the $25.6M Arup case)

Business email compromise grew a face. In the Arup case, reported by CNN in 2024, a Hong Kong finance employee got a message from the "UK CFO" about a confidential transaction. He suspected phishing, until he joined a video call with the CFO and several colleagues. Every participant except him was a deepfake built from public footage. He approved 15 transfers totaling $25.6 million (HK$200 million).

The FBI's 2025 IC3 report puts business email compromise losses at $3 billion for the year, and deepfakes are now the closing tool for the biggest BEC plays. Finance teams need callback rules for any payment instruction, no matter who appears on screen. Full guide: executive deepfake fraud.

2. Voice cloning and family emergency scams

The consumer-side twin of CEO fraud. Scammers clone a family member's voice from social media audio, then call with a staged emergency: a kidnapping, an arrest, a crash. In 2023, Arizona mother Jennifer DeStefano heard her daughter's cloned voice before a man demanded $1 million; she later testified before the Senate Judiciary Committee. McAfee's 2023 research showed three seconds of audio can produce an 85 percent voice match.

Defense comes down to a family code word and a callback rule. Full guide with documented cases and a printable family playbook: voice cloning scams.

3. Grandparent scams with AI voices

A targeted variant of voice cloning aimed at older adults, and the cruelest one. The "grandchild" calls in trouble, needing bail or hospital money, begging grandma not to tell anyone. In 2025, CBC reported an Ontario senior who nearly paid after a caller used what police suspected was an AI clone of her grandson's voice claiming he was under arrest.

The FBI's 2025 IC3 report shows why this segment matters: Americans 60 and older reported $7.7 billion in fraud losses in 2025, up about 60 percent year over year. Full guide: grandparent scams with AI voices.

4. Deepfake investment scams and fake celebrity endorsements

Deepfaked celebrities pitch trading platforms and crypto schemes they have never heard of. Elon Musk, news anchors, and finance personalities are the favorite faces; Resemble AI's Q1 2025 Deepfake Incident Report found public figures were impersonated in 47 percent of documented deepfake incidents. The videos run as paid social ads, lending stolen credibility to pig butchering and fake trading apps.

The category money is enormous: the FBI's 2025 IC3 report attributes $8.6 billion in losses to investment fraud, with cryptocurrency-related fraud reaching $11.3 billion. The rule: no real celebrity invites you to an investment platform in a targeted ad. Full guide: deepfake investment scams.

5. AI romance scams and deepfake catfishing

Romance scammers now back their fake personas with AI-generated profile photos, cloned voices, and even live face-swapped video calls. In January 2025, a French woman revealed she had sent roughly $850,000 to scammers who courted her for months using AI-generated images of "Brad Pitt," including fake hospital photos.

The new playbook means "we video chatted, so he must be real" no longer holds. Verify identity through multiple channels and reverse-search profile images before trust deepens. Full guide: AI romance scams.

6. Deepfake phishing and vishing at work

Below the headline-grabbing CEO frauds sits a steady stream of deepfake phishing: cloned executive voicemails asking for gift cards, fake IT support calls harvesting credentials, video messages nudging employees toward malicious links. In 2024, scammers cloned the voice of Ferrari's CEO in WhatsApp calls to an executive, who foiled the attempt by asking a question only the real CEO could answer.

Pindrop's 2025 Voice Intelligence and Security Report measured a more than 1,300 percent surge in deepfake fraud attempts at contact centers in 2024, from roughly one per month to seven per day. Train teams to verify any unusual request through a second channel, every time. Full guide: deepfake phishing.

7. KYC and identity verification fraud

Not all deepfake scams target people. Many target systems: fraudsters use face swaps and injection attacks to pass Know Your Customer onboarding at banks, exchanges, and fintechs. Entrust's 2025 Identity Fraud Report found a deepfake attempt occurred every five minutes in 2024, and deepfakes made up 24 percent of fraudulent attempts against motion-based biometric checks.

For businesses, the fix is layered: liveness detection, injection-attack monitoring, and document forensics. Full guide: KYC deepfake detection.

8. Video call impersonation on Zoom and Teams

Real-time face swap tools now run live on video calls. Targets include executives (the Arup pattern), job interviews, and even government officials: in September 2024, someone posing as a former Ukrainian foreign minister held a deepfake video call with US Senator Ben Cardin before staff flagged inconsistencies.

Treat the meeting platform as untrusted. Ask the person to turn their head, wave a hand in front of their face, or answer a question outside the script. Then verify on a second channel before acting. Full guide: deepfake video call scams.

• 2019A cloned-voice call steals about $243,000 from a UK energy company.
• 2023A faked kidnapping call using an AI voice targets the DeStefano family.
• 2024Arup loses $25.6M to a deepfaked video call impersonating executives.
• 2025The FBI logs AI-assisted cybercrime for the first time, with nearly $900M in losses.

How to Spot a Deepfake Scam in the Moment

Every type above shares the same behavioral fingerprint. Run this checklist when anything urgent arrives by phone, video, or DM:

• Urgency plus secrecy. "Now" and "tell no one" together are a near-certain scam signal.
• Channel switching. The contact starts on one platform and pushes you to another, away from records and safety checks.
• Payment red flags. Wire transfers, gift cards, crypto, cash couriers. Legitimate institutions do not collect this way.
• Verification resistance. Any excuse why you cannot call back, ask questions, or loop in a colleague.
• Too-perfect or too-glitchy media. Flat audio emotion, mismatched lip sync, odd lighting, faces that blur on fast movement.

For media-specific tells, see our guides on how to detect a deepfake and how to detect deepfake audio.

How to Verify Suspicious Media With a Deepfake Detector

When you have the actual file, a voicemail, a video, a profile photo, verification takes under a minute:

1. Save the media. Download the video, audio, or image rather than judging it in-app.
2. Upload it to DeepfakeDetector.ai. We analyze video, image, and audio with 95 percent accuracy.
3. Read the verdict. You get an AI or Human verdict with a TrustScore from 0 to 100. Treat a low TrustScore as a stop sign: verify through a second channel before any money or credentials move.

The free plan includes 50 detections per month, with video and voice clips up to 2 minutes each. No detector is 100 percent, ours included, so combine the scan with the behavioral checklist above.

CTA: Check any suspicious video, image, or audio free: 50 detections per month, files deleted within 60 seconds of analysis. Create a free account

Latest Deepfake Scam News and Cases

Updated June 2026. This section is refreshed monthly.

• June 2026: The FBI issued a renewed warning about AI voice cloning scams that mimic loved ones in distress, citing continued growth in family emergency schemes.
• May 2026: US enforcement of the TAKE IT DOWN Act began, allowing fines of over $53,000 per non-consensual deepfake image platforms fail to remove within 48 hours.
• April 2026 (reporting on 2025 data): The FBI's 2025 IC3 report recorded its first AI-assisted cybercrime category: 22,000+ complaints, nearly $900 million in losses.
• 2025: Resemble AI documented 163 deepfake incidents in Q1 2025 alone, with more than $200 million in fraud losses in the quarter.
• 2025: Sumsub's Identity Fraud Report recorded a 1,100 percent surge in deepfake fraud in the US in Q1 2025, with deepfakes reaching 7 percent of all fraudulent activity globally.

FAQ

What is the most common deepfake scam? Voice-based impostor scams lead by volume. The FTC logged impostor scams as the most reported fraud category of 2025, with losses above $3.5 billion, and AI voice cloning increasingly powers them.

How much money is lost to deepfake scams? No single total exists, but the markers are clear: $25.6 million in the 2024 Arup case alone, $200 million+ documented by Resemble AI in Q1 2025, and nearly $900 million in AI-assisted cybercrime complaints in the FBI's 2025 IC3 report. Our deepfake statistics page tracks every sourced figure.

Can deepfake scams be detected? Yes, with layers. Behavioral red flags (urgency, secrecy, untraceable payment) catch most attempts, and detection tools score the media itself. No method is 100 percent, so verify through a second channel before acting.

What should I do if I was targeted? Stop contact, secure your accounts, contact your bank if money moved, and report to the FTC at reportfraud.ftc.gov and the FBI at ic3.gov. Our response playbook covers the first 48 hours step by step.

Are deepfake scams illegal? Yes. Fraud laws apply regardless of the technology, and deepfake-specific laws are spreading fast. See our guide to whether deepfakes are illegal.

Conclusion: Verify First, Act Second

Deepfake scams come in eight flavors, but they share one engine: borrowed trust under time pressure. The defense is the same everywhere. Slow down, verify the person through a channel you control, and verify the media with a detector before money, credentials, or secrets move. Explore the deep-dive guides above for each scam type, and keep this page bookmarked as new cases land.

CTA: Verify before you trust. Run any suspicious video, image, or audio through DeepfakeDetector.ai free, 50 detections per month. Get started free

Related reading