Deepfake Statistics 2024

Global deepfake fraud losses, frequency data, corporate impact, and AI image detection accuracy by model. Sources include FBI IC3, FTC, Medius, and McAfee.

Sources: FBI IC3 • FTC • Medius Group • McAfee • Updated July 2026

Every 5 min
Deepfake fraud attempt (2024)
+118%
YoY increase in deepfake fraud (2023→2024)
$12B
Global deepfake fraud losses (2024)
90%
US companies targeted by deepfake fraud

Scale of deepfake fraud in 2024

A deepfake fraud attempt occurs approximately every 5 minutes globally. The annual frequency increased 118% from 2023 to 2024, and since 2022 — when tools like Midjourney, Stable Diffusion, and DALL-E 3 became widely accessible — deepfake fraud incidents have grown over 1,700%.

Global deepfake-related fraud losses reached an estimated $12 billion in 2024 and are projected to reach $40 billion by 2027 as AI generation tools become cheaper and detection difficulty increases. This growth is driven primarily by three use cases: romance scam profile photos, corporate fraud (fake executive video calls), and financial institution identity bypass.

The FBI's 2024 IC3 report confirmed that romance scammers are actively using AI-generated images in their operations. The FTC reported $12.5 billion in total fraud losses in 2024 (+25% vs. 2023), with a growing proportion attributed to AI-assisted deception. Pig butchering scams — which combine fake AI-generated romantic personas with cryptocurrency investment fraud — accounted for $5.7 billion in FTC-reported losses alone.

Sources: FBI IC3 2024 Annual Report • FTC Consumer Sentinel 2024 • Synthetic Media Fraud Research 2024

Detect deepfakes in any photo — right now

🕵️ Add to Chrome — Free

10 checks/day free · No account required · Works on any website

Corporate impact: who deepfakes target

A Medius Group survey of 1,500 US finance executives in 2024 found that 90% of their companies had experienced deepfake fraud — yet most lacked specific detection protocols. Separately, 49% of companies reported being targeted by audio or video deepfake attacks in 2024 (enterprise security industry data).

The most common corporate deepfake attacks are:

  • Fake CEO video calls — A notable Hong Kong case in early 2024 resulted in a $25 million loss after an employee was deceived by a deepfaked video conference with an AI-generated CFO and colleagues.
  • AI voice cloning — AI can clone a person's voice from as little as 3 seconds of audio (McAfee 2024). 70% of people cannot distinguish an AI-cloned voice from the real person (McAfee consumer survey).
  • Identity document fraud — AI-generated faces are being used to bypass KYC (know your customer) systems at financial institutions. Biometric fraud via AI grew 704% from 2022 to 2023.
  • Deepfake political content — 77% of US voters encountered at least one deepfake related to the 2024 presidential election (2024 election integrity research).

Sources: Medius Group Finance Deepfake Report 2024 • McAfee Consumer Survey 2024 • ITRC 2024 • Enterprise Security Magazine

Deepfake detection accuracy by AI model (2024–2026)

Detection accuracy varies dramatically by AI image generator. Open-source detection models have struggled to keep pace with advances in AI image generation. Human eyes perform even worse — studies consistently show people can identify deepfakes at rates only slightly better than random chance.

AI Generator Open-Source Detection Hive AI (Faux Spy) Why It's Hard
Adobe Firefly v4 ~18% ~94% Enterprise-grade rendering; trained on licensed stock photos
Flux Dev (FLUX.1) ~21% ~94% Open-source; fastest-growing generator 2024–26
DALL-E 3 / ChatGPT ~31% ~94% Mass reach via ChatGPT; advanced coherence
Google Imagen 4 ~19% ~94% Google's proprietary training data advantage
Midjourney v7 ~24% ~94% Photorealistic mode eliminates most artifacts
Stable Diffusion v1.4 ~73% ~94% Older model; more detectable artifacts remain

Open-source accuracy figures represent published researcher benchmarks (2024–2026). Hive AI accuracy reflects Faux Spy's internal testing across generator models. Accuracy varies by image type and model version.

Deepfakes and romance scams

Deepfake technology has fundamentally changed how romance scammers operate. Where scammers once had to steal real photos from social media — images that could be caught by reverse image search — they now generate entirely fictional faces. A synthetic face has no real-world identity to trace, making it invisible to traditional detection methods.

Key data points on the deepfake-romance scam intersection:

  • $672 million — Romance and confidence fraud losses reported to the FBI IC3 in 2024, across 17,910 complaints
  • $1.14 billion — Romance scam losses reported to the FTC in 2023 (most recent full year)
  • $4.7 billion — Estimated true annual losses including unreported cases (Consumer Federation of America)
  • 1 in 4 Americans — Encountered an AI-generated or AI-modified photo on a dating app in 2026 (McAfee)
  • 55% of malicious clones on Tinder now use AI-generated images (McAfee 2026)
  • FBI confirmation — The FBI's 2024 IC3 annual report explicitly identified AI-generated images as an escalating tool in romance scam operations

See full data: Romance Scam Statistics 2024Fake Dating Profile Statistics

Sources: FBI IC3 2024 Annual Report • FTC Consumer Sentinel 2024 • McAfee Deepfakes and Dating 2026 • Consumer Federation of America

Deepfake statistics: frequently asked questions

How common are deepfake fraud attempts?

A deepfake fraud attempt occurs approximately every 5 minutes globally (2024 data). The annual frequency grew 118% from 2023 to 2024, and has increased over 1,700% since 2022 when AI generation tools became widely accessible to the public.

How much money do deepfakes cost in fraud each year?

Global deepfake-related fraud losses reached an estimated $12 billion in 2024. This is projected to grow to $40 billion by 2027 as AI image and video generation becomes cheaper and more convincing. The $12B figure includes romance scams, corporate fraud, and financial institution fraud — but not the full cost of misinformation and political deepfakes.

What percentage of companies have been targeted by deepfake fraud?

90% of US companies experienced deepfake fraud in 2024, per a Medius Group survey of 1,500 finance executives. In a separate measure, 49% of companies reported being targeted by audio or video deepfake attacks specifically. Corporate deepfakes most commonly impersonate executives to authorize fraudulent wire transfers.

Can people tell the difference between real and AI-generated images?

No — not reliably. 68% of deepfakes are now nearly indistinguishable from real images by human perception (2024 benchmark research). In audio deepfakes, 70% of people cannot tell the difference between a real voice and an AI clone (McAfee 2024). AI voice cloning requires as little as 3 seconds of source audio to produce convincing results.

Which AI image generators are hardest to detect?

Adobe Firefly v4 is the hardest to detect at ~18% accuracy using open-source detection models. Flux Dev (~21%), Google Imagen 4 (~19%), and Midjourney v7 (~24%) are similarly difficult. Commercial detection models like Hive AI — which powers Faux Spy — achieve approximately 94% accuracy across all of these generators.

How do you detect a deepfake image?

The most reliable method is a commercial AI detection tool. Open-source models detect newer AI generators at rates of only 18–31%, not much better than guessing. Commercial models like Hive AI achieve ~94% overall accuracy. Faux Spy is a free browser extension powered by Hive AI that lets you right-click any photo on any website and get an instant AI vs. Real verdict with a confidence score — no uploading required.

Are deepfakes illegal?

It depends on how they're used. Creating a deepfake is not inherently illegal in most jurisdictions. However, using a deepfake to commit fraud, impersonate someone without consent, create non-consensual intimate imagery, or interfere with elections is illegal in most US states and many countries. As of 2025, 47 US states have passed or are advancing deepfake-specific legislation.

How has deepfake fraud changed since 2022?

Deepfake fraud has grown over 1,700% since 2022. The annual frequency rose 118% from 2023 to 2024 alone. The growth is directly tied to public availability of AI image generators: Midjourney launched publicly in 2022, Stable Diffusion became open-source in 2022, and DALL-E 3 was integrated into ChatGPT in 2023 — bringing advanced AI image generation to hundreds of millions of people at no cost.

Detect AI-generated images in seconds

Faux Spy is the only browser extension that scans photos for AI generation and deepfake manipulation directly on dating apps, social media, and any other website — no uploading required.

🕵️ Add to Chrome — Free 🦊 Add to Firefox

10 checks/day free · No account required · 94% accuracy (Hive AI)

Related statistics