An ad fraud firm says AI played a role in 52% of the scam schemes it found this year, and the ads now look far more convincing.
AI was involved in just over half of the online ad fraud schemes that one security company detected between January and September 2026. The firm, ADEX, put the number at 52% in a new report. The scams are not new, but they now look polished enough that even people who know AI may not spot them right away.
ADEX sells anti-fraud tools to advertising platforms. Its report draws on moderation data from eight platforms that use its service, covering 1.2 million campaigns and 4.9 million ad creatives. Among the AI-linked schemes, bots posing as real people showed up in 64%. Cloned brands or fake endorsements appeared in 53%. Deepfakes appeared in 11%. These groups overlap, because one scheme often uses several tricks at once.
A couple of years ago, a fake investment ad often meant a rushed web page and a celebrity photo with a made-up quote. A small team then chatted with victims in private messages. Today, ADEX says, AI can copy a news site or a financial institution almost perfectly. A chatbot can then hold conversations around the clock, in the victim's own language.
Sergey Martianov, ADEX's Chief Product Officer, said the basic scheme has not changed: borrow a trusted face, promise big returns, move the victim into a private chat. What changed is the finish. The report says brand cloning shows up in investment and blockchain scams, fake online stores, and fake security alerts that imitate well-known tech brands.
Speed is another problem. ADEX saw cases where a rejected scam ad came back in a revised form within hours, sometimes the same day. The firm says the turnaround varies a lot and gives no standard time.
ADEX expects scammers to start targeting AI shopping assistants in 2027. These are tools that browse sites, compare products and buy things for people. One method is called prompt injection, where hidden instructions are placed in content an AI tool reads, such as invisible page elements or product images. Scammers could also show one page to an AI agent and another to a human.
This is a forecast. ADEX says it is not an attack type it currently measures in its data. It does point out that prompt injection is already described in security research.
Anyone who sees online ads is exposed, especially to ads for investments, crypto and online stores. Businesses face a different risk, since scammers can copy their brand name in ads. Keep in mind that the figures come from a company that sells fraud protection, and they count schemes, not individual ads or accounts.
The alert box above lists the steps for readers and business owners. Business owners should also ask their technology partners how they tell a helpful automated shopper from a harmful one. ADEX suggests three tests: can the visitor prove who it is, does it act in line with its stated task, and do its sessions end in something real, like a purchase or a booking.
A convincing ad is no proof of a real company, so check the web address and be wary of big return promises before you click, invest or buy.
A viral ChatGPT prompt turns what the chatbot remembers about you into a room image. Posting it publicly could give scammers clues to target you.
Read more: Viral ChatGPT room image could hand scammers clues about youFake PayPal emails and texts, some posing as a security report, try to get you to tap a link and type your login. Checking your account directly avoids the trap.
Read more: Got a fake PayPal message? Skip its link and check your account yourselfSecurity firm Flashpoint says criminals are trading in refund tricks, gift card fraud and hacked online stores ahead of the holidays. Shop owners should check their refund steps.
Read more: Scammers are targeting store refunds and gift cards before the holidays
