AI Scams in 2026: All 12 Types Explained, With Real Examples

---

Americans reported $12.5 billion in fraud losses in 2024, a 25 percent jump in one year, per the FTC's March 2025 data release. The FBI now tracks why: its 2025 IC3 Annual Report logged more than 22,000 complaints and nearly $900 million in losses tied to AI-assisted cybercrime, the first year the bureau counted AI as a crime descriptor. This guide covers every major AI scam on one page: 12 types, five channels, real cases, and the habits that beat them.

What is an AI scam? An AI scam is any fraud that uses artificial intelligence, such as cloned voices, deepfake videos, AI-written phishing, or chatbot impersonation, to deceive victims. The 12 major types fall into five channels: calls, video, text, images, and ads. Verification habits and detection tools stop most of them.

What Is an AI Scam?

An AI scam is a con where artificial intelligence does work a human scammer used to do: writing the message, faking the voice, generating the face, or holding the conversation. The fraud underneath is old. The impostor scam, the romance scam, and the fake invoice all predate computers.

What changed is the economics. AI lets one scammer run thousands of personalized attacks at once, in fluent English, with audio and video that pass casual inspection. The FTC warned as early as 2023 that scammers were cloning family members' voices from short online clips to fake emergencies.

Three things make artificial intelligence scams harder to spot than their ancestors:

• Scale: one operator, thousands of simultaneous targets.
• Personalization: messages reference your real job, family, or posts.
• Believability: the old tells, like typos and robotic voices, are gone.

The loss data behind every claim in this guide lives on our deepfake statistics page, each figure linked to its primary source.

The 12 Types of AI Scams (Organized by Channel)

The fastest way to understand AI scams is by the channel they arrive on. Here are all 12 types:

Calls: AI phone scams

1. AI voice cloning scams. Three seconds of audio is enough to clone a voice with an 85 percent match, per McAfee's Beware the Artificial Impostor report (2023). Scammers pull that audio from social videos and voicemail greetings, then call posing as your child, spouse, or boss with an urgent money request. The defense is simple: hang up and call the person back on the number you already have. Our full guide to voice cloning scams covers the documented cases.

2. Grandparent scams, upgraded. The classic "Grandma, I'm in jail" call now arrives in the grandchild's actual voice. The FTC's impersonation data spotlight (2025) found reports of older adults losing $10,000 or more to impersonation scams more than quadrupled from 2020 to 2024. A family code word, agreed in advance and never posted online, defeats the clone.

3. AI robocalls. Generated voices now deliver robocalls that sound like banks, utilities, and even politicians. After a fake presidential voice targeted New Hampshire voters in January 2024, the FCC ruled AI-generated voices in robocalls illegal under the Telephone Consumer Protection Act. Treat any unsolicited call asking you to press a button, share a code, or confirm account details as hostile, however human it sounds.

Video: deepfake scams

4. Deepfake video call fraud. The landmark case: an employee at engineering firm Arup wired $25.6 million after a video meeting where every other participant, including the CFO, was a deepfake, per CNN (2024). We break down the full timeline in our guide to executive deepfake fraud. Seeing a familiar face on a live call is no longer proof of identity.

5. Fake celebrity investment videos. Deepfaked clips of Elon Musk, national news anchors, and finance celebrities push fraudulent trading platforms and crypto giveaways. The categories where these videos concentrate are enormous: investment fraud cost $8.6 billion and crypto-related fraud $11.3 billion in 2025, per the FBI IC3. No legitimate figure announces giveaways through ads. Our deepfake scams pillar maps every video variant.

Text: AI-written scams

6. AI-written phishing emails. The FBI warned in a December 2024 public service announcement that criminals use generative AI to mass-produce phishing emails without the spelling errors that once gave them away. Here is the honest truth: you usually cannot reliably tell AI-written text from human text. So judge the request, not the prose. Any message that asks for credentials, payment, or urgency deserves verification through a separate channel.

7. Smishing at scale. AI lets fraud rings blast convincing fake texts about unpaid tolls, stuck packages, and frozen bank accounts, each variant A/B tested like marketing copy. A related classic still circulating is the Google Voice verification code scam, where a "buyer" asks you to read back a code that hijacks your number. Never share a verification code with anyone who contacted you first.

8. Chatbot romance scams. Romance scammers now use chatbots to keep dozens of victims in warm, attentive conversations around the clock. Reported romance scam losses already ran to $1.14 billion in 2023, per the FTC's data spotlight, before chatbots made the workload trivial. Watch for partners who chat endlessly but always dodge live video. Our guide to AI romance scams covers the warning signs in depth.

Images: AI-generated photos and documents

9. Fake profile pictures and catfishing. AI face generators produce unique, realistic photos that defeat reverse image search, the classic catfish check. These faces front fake recruiters, fake investors, and fake romantic interests. Humans are poor judges here: only 0.1 percent of participants in an iProov study (2025) correctly identified all the real and AI-generated media shown to them. Verify people through video calls plus independent channels, not photos.

10. AI-generated document fraud. Fake IDs, bank statements, invoices, and insurance claims are now generated on demand. Digital document forgeries rose 244 percent year over year, per the Entrust 2025 Identity Fraud Report (2024), and synthetic identity document fraud jumped more than 300 percent in the US, per Sumsub (2025). Businesses should verify documents against issuing sources, not appearances.

Ads and listings: AI scams in your feed

11. Deepfake endorsement ads. Paid social ads featuring a deepfaked celebrity face or voice endorse miracle supplements, trading apps, and giveaways. The ad format buys borrowed trust, and platforms remove these slower than scammers re-upload them. Before acting on any endorsement, check the person's official site or verified accounts. If the product is real, it will be announced there too.

12. AI-generated fake stores and reviews. Entire storefronts are now AI-built: product photos, descriptions, support chat, and glowing reviews, none of them real. The FTC's rule banning fake reviews and testimonials (2024) explicitly covers AI-generated reviews. Protect yourself with payment rails that offer chargebacks, a domain age check, and a search for the store's name plus the word "scam" before buying.

How to Avoid AI Scams: 6 Habits That Work

You do not need to out-tech the scammers. These six habits defeat nearly every type above:

1. Verify on a second channel. Whatever arrived by call, text, or video, confirm it by contacting the person on a number or address you already had.
2. Set a code word. Agree on a private word with family, and challenge phrases with coworkers, that no clone can know.
3. Slow down on urgency. Every AI scam manufactures time pressure. Real institutions let you call back.
4. Treat the payment method as the tell. Gift cards, wire transfers, and crypto are the currencies of fraud. Legitimate organizations do not demand them.
5. Starve the clone. Limit public audio and video of yourself and your kids, and set social profiles to private where you can.
6. Verify the media itself. When a video, photo, or voice note drives the request, run it through a deepfake detector before you act.

How to Check Suspicious Media for AI Manipulation

Habits cover the request. A detector covers the file. Upload the suspicious video, image, or voice note to DeepfakeDetector.ai and you get a clear verdict in seconds: Authentic, Likely Synthetic, or Inconclusive, plus a TrustScore from 0 to 100 showing how confident the analysis is.

The platform detects AI-generated video, image, and audio with 95 percent accuracy, and files are deleted from primary storage within 60 seconds of analysis unless you opt to retain them. The free plan includes 50 detections per month, enough to check every suspicious file that reaches your family.

CTA: Suspicious video, photo, or voice note? Verify it free, 50 detections per month. Create a free account

FAQ

What are the most common AI scams right now? Impostor scams lead. They were the most reported fraud category with the FTC, and AI voice cloning and AI-written phishing are their growth engines. The FTC's 2025 impersonation spotlight shows large-loss impersonation reports more than quadrupling since 2020.

How can you tell if a message was written by AI? You usually cannot, and tools that claim otherwise are unreliable. Judge the request instead of the prose: unexpected contact, urgency, secrecy, and unusual payment methods are the real tells, whoever or whatever wrote the words.

Are AI scams illegal? Yes. Fraud is illegal regardless of the tools used, and AI-specific rules are stacking on top, like the FCC's ban on AI voices in robocalls. Several states have added their own statutes. See our guide to whether deepfakes are illegal.

Where do I report an AI scam? If money moved, call your bank first and ask for a recall. Then report to the FTC at reportfraud.ftc.gov and the FBI at ic3.gov. Reports build the loss data that drives enforcement.

Do AI scam detection tools work? Media detectors work well on video, images, and audio. Machines detect manipulation artifacts that humans miss, which matters when studies show people catch only a fraction of quality fakes. Pair a detector with the verification habits above, since no tool checks a scammer's intentions.

Conclusion: New Tools, Same Con

Every AI scam in this guide runs on the same two-stroke engine that powered cons a century ago: borrowed trust plus manufactured urgency. The AI only makes the borrowing easier. So the defense stays human and boring: verify people on channels you control, never pay under pressure, and check suspicious media before acting on it. Bookmark this page, share it with the people most likely to get the call, and keep our deepfake scams pillar handy for the video-specific variants.

CTA: Check any suspicious video, image, or voice note free, 50 detections per month. Get started free

Related reading