Scams have always relied on persuasion: a convincing story, a sense of urgency and a trusted voice. AI has made all three cheaper and more convincing. Understanding the main tactics, and the simple habits that defeat them, is the best defense.
Updated September 2026: we added links to warnings from the FTC, FCC and FBI, the latest US fraud figures, and sources for the cases described.
How big the problem is
People in the US reported losing about $16 billion to fraud in 2025, up 25% on the year before, according to Federal Trade Commission data. Impostor scams, where someone pretends to be a person or organization you trust, were the most commonly reported type, accounting for nearly one in three reports and $3.5 billion in losses. The FTC’s figures don’t say how many of those scams used AI, but the FBI has warned that criminals use generative AI to make every stage of these schemes more believable.
Tactic 1: Voice cloning
With a short audio clip, which can come from social media videos or voicemail greetings, readily available tools can produce a convincing imitation of someone’s voice. The FTC warned in 2023 that scammers need only “a short audio clip of your family member’s voice” to stage a fake emergency call. Scammers also impersonate managers to ask employees for urgent transfers.
In February 2024 the Federal Communications Commission ruled that AI-generated voices in robocalls count as "artificial" under the Telephone Consumer Protection Act, which makes most such calls illegal without prior consent.
Tactic 2: Deepfake video calls
In early 2024 a finance employee at the Hong Kong office of engineering firm Arup transferred HK$200 million, about $25 million, after a video call in which the company’s chief financial officer and other colleagues were deepfakes. Arup confirmed in May 2024 that it had been the target. Business email compromise and executive impersonation scams now increasingly include voice and video.
Tactic 3: Flawless phishing
Bad spelling used to be a giveaway. AI writes fluent, personalized messages in any language, referencing details scraped from public profiles. The FBI’s December 2024 alert notes that generative AI helps criminals avoid the grammar and spelling mistakes that once exposed fraud. Fake customer support chats and cloned websites are also easier to produce.
Tactic 4: Fake investment and romance schemes
AI-generated images, fake celebrity endorsement videos and chatbots that keep conversations going at scale are used to build trust over weeks before asking for money, often in cryptocurrency. The FBI says criminals use generated images to create fake social media profiles and fraudulent identification documents. See our personal finance guide for the red flags regulators highlight in investment offers.
Warning signs that still work
- Urgency and secrecy: “Do not tell anyone, it has to be now.”
- Unusual payment methods: gift cards, crypto, wire transfers to new accounts. The FTC lists all three as common signs of a scam.
- Requests that bypass normal processes, especially at work.
- Emotional pressure, such as panic, fear or excitement.
- Contact through an unexpected channel or a new number.
Three habits that stop most AI scams 1. Agree a family code word for emergencies, as the FBI recommends. 2. Hang up and call back on a number you already know. 3. Never move money based on a single call, message or video, however real it seems.
For businesses
- Require call-back verification on a known number for any payment change or urgent transfer request.
- Use multi-person approval for large payments.
- Train staff with real examples of voice and video impersonation, such as the Arup case.
- Limit publicly available recordings of executives where practical.
- Set up a clear, blame-free way to report suspicious requests.
If you are targeted
Stop contact, do not send money and verify independently. If you have already paid, contact your bank immediately, since speed matters for recovering funds. In the US, report the scam to the FTC at ReportFraud.ftc.gov and to the FBI’s Internet Crime Complaint Center; elsewhere, contact the police and your national fraud reporting service.
Verify, don’t judge
AI makes fakes more convincing, so the defense has to move from spotting flaws to verifying through trusted channels. As our deepfakes guide explains, process beats perception. A simple rule, never act on urgency alone, protects against the newest scams just as it did against the oldest.
Sources
- Scammers use AI to enhance their family emergency schemes, Federal Trade Commission, March 2023
- FCC Makes AI-Generated Voices in Robocalls Illegal, Federal Communications Commission, February 2024
- Criminals Use Generative Artificial Intelligence to Facilitate Financial Fraud, FBI Internet Crime Complaint Center, December 2024
- Incident 634: Deepfake CFO scam costs Arup $25 million, AI Incident Database (compiling CNN and other reporting)
- FTC Data Show People Reported Losing $3.5 Billion to Imposter Scams in 2025, Federal Trade Commission, June 2026
