Researchers found a network of 100+ polished look‑alike subscription websites that impersonate real products (and invent new ones) to sell expensive “AI” plans. The sites rely on professional landing pages and a real Google sign-in flow to appear legitimate, then steer visitors into paid subscriptions and file uploads despite unclear ownership and support details.
Researchers identified more than 100 subscription websites linked by identical underlying files and closely related developer email addresses, suggesting a single operator or connected group runs many of them. Some sites impersonate well-known products, including GPT-6 Astra, DaVinci Resolve, PixAI, OpenCut, and Omegle. Rather than relying on fake password forms or malware, the sites use polished product pages and a genuine Google sign-in flow to build trust. After signing in, visitors are shown paid plans, usage credits, and payment histories, with pricing ranging from less than $10 a month to more than $2,000 a year.
The scam works because the visible signals people are trained to trust, such as HTTPS padlocks, professional design, and confident marketing language, are all present. The Google sign-in step adds a further layer of false confidence: because the password is entered on Google's own site, visitors assume the underlying service must also be legitimate. In reality, a genuine sign-in page only confirms the connection is encrypted and that Google authenticated the login, not that the branded service is who it claims to be. Consent screens in examined cases listed support contacts using free webmail addresses rather than addresses tied to the displayed brand.
Employees across marketing, creative, procurement, and finance roles are common targets for this pattern since they routinely evaluate and purchase software tools. Encourage staff to check consent screen details against the intended brand before approving access, to look for independently verifiable company information before subscribing, and to avoid annual commitments or file uploads to services that cannot demonstrate legitimacy. IT helpdesk teams can support this by giving simple guidance on reading OAuth consent screens and reporting suspicious subscription offers before purchases go through, particularly in expense approval and vendor management workflows (see attack.mitre.org/techniques/T1598 for related reconnaissance techniques and attack.mitre.org/techniques/T1204/001 for the malicious link execution pattern).
Mirage safely runs attacks like this one against your own team, so you find out what happens before a real adversary does.
The sites send visitors to a genuine Google sign-in page where the password is entered on Google's own website, but a real sign-in page does not confirm the service itself is official or connected to the brand it displays.
Researchers found sites impersonating well-known products and brands including GPT-6 Astra, DaVinci Resolve, PixAI, OpenCut, and Omegle.
Warning signs include a Google consent screen listing a free webmail support address instead of a company address, no way to test the product before paying, and no verifiable business name or registered address.
Several of the sites ask users to upload documents or recordings for processing, meaning users may share account information and sensitive material with an operator whose identity they cannot confirm.
You land on “GPT-6 Astra” or a shiny DaVinci Resolve AI page, hit Get started, and it looks totally legit. Behind that Get started button is a subscription service. These sites send you to a genuine Google sign‑in page, then hit you with plans from under ten bucks a month to over two thousand a year. Here’s the catch: a real Google sign‑in doesn’t prove the service is real. The consent screen lists a random free webmail support address, no real company name, no trial, but it wants your Google profile, your card, and even your file uploads. Your move: before you pay or upload anything, pause on the Google consent screen and check who actually owns the app. If the developer email or company name doesn’t match the brand you expect, close the tab.
Criminal groups are stealing Meta Business Manager and Google Ads accounts using phishing that arrives through trusted platforms like Salesforce, Google Workspace mail-merge, and SharePoint links. The stolen accounts are valuable not just for the budget inside them, but because older accounts with…
A phishing email posing as a ChatGPT billing notice pressures users to “update payment information” to avoid service interruption. The button routes through a Google redirect and lands on a fake OpenAI login page that captures usernames and passwords before sending victims to an error page.
Cofense observed a real phishing email that impersonates OpenAI/ChatGPT billing to trick users into “updating” payment details. The email uses the real ChatGPT logo, urgency (“48 hours”), and a prominent button to drive clicks to a lookalike ChatGPT login page. Any credentials entered are harvested…
A large phishing (smishing) campaign is sending messages that claim a T-Mobile customer’s rewards points are about to expire. The texts use urgency, made-up point balances, and lookalike “t-mobile.*.top” links to push people into clicking and entering sensitive information. Malwarebytes observed…
Check Point researchers demonstrated a real proof-of-concept where a victim’s ChatGPT session could be tricked into running hidden, attacker-controlled tasks in parallel with the user’s normal request. In the demo, the attacker used a covert cross-account channel to make ChatGPT access the victim’s…
Researchers describe a phishing workflow where an attacker logs into a victim’s Google account using stolen password + authenticator code, then quickly enrolls a new passkey to keep access even if the password is changed. The trick relies on victims choosing a weaker sign-in fallback (one-time…
