Security
Island links 186 lookalike domains impersonating 27 AI brands to ClickFix lures, browser malware and search-hijacking extensions, in a campaign it calls TypoPrompting
Island found 186 lookalike domains copying 27 AI brands; mistyped names lead to ClickFix lures, browser malware and search hijackers. No victims were reported.
Browser-security company Island has inventoried 186 lookalike domains that imitate 27 AI brands, including Claude, ChatGPT, Gemini and DeepSeek, and says mistyped addresses on some of them lead to ClickFix lures, browser-hijacking malware and search-hijacking Chrome extensions. It calls the pattern TypoPrompting. The 186 figure is an inventory of domains, not confirmed infections, and no victims or losses were reported in the coverage read.
The essentials
- The inventory: 186 lookalike domains across 27 AI brands. ChatGPT is the most imitated, with 25 domains, per Cryptopolitan. Examples include cluade[.]ai, chatlgpt[.]com and preplexity[.]ai. Some are only parked pages.
- ClickFix route: a parked page redirects through a chain that fingerprints the browser, then shows a fake Cloudflare verification page that gives the user a PowerShell command to run. The Hacker News also describes macOS and archive variants.
- Net Guard route: fake security-check pages deliver Net Guard, a Microsoft Store app posing as a Wi-Fi and DNS scanner. Island says the download is identical to the Store version, and that once installed it can change the default search engine and install Chrome extensions.
- Search-hijack route: one lookalike redirects to the Quick Search Chrome extension, which overrides the default search provider. Island counts 16 live Chrome Web Store listings with 972,000 declared users as of November 3.
Declared users are not victims
The 972,000 figure is the total of users shown on the extension listings, not infections caused by this campaign. The coverage read reports no confirmed victims.
Why typos are the entry point
The campaign relies on people mistyping a familiar AI name. Island’s closing point is that reaching the right AI service should be a managed path, not a test of typing accuracy. A related fingerprinting script was also found on 211 entry domains, including lookalikes of banks and enterprise software.
What is not established
- How many people visited these domains or were infected: not reported.
- Who operates the campaign: not established in the sources read.
- Whether all 186 domains are malicious: Island’s inventory includes parked domains.
- Whether the Chrome extensions have since been removed: not stated.
Summary table
| Point | What is stated as of November 7, 2026 | Source |
|---|---|---|
| Domains | 186 lookalikes across 27 AI brands | Island |
| Most imitated | ChatGPT, with 25 domains | Cryptopolitan |
| ClickFix | Fake Cloudflare check serves a PowerShell command | The Hacker News |
| Net Guard | Microsoft Store app posing as a Wi-Fi and DNS scanner | The Hacker News |
| Extensions | 16 Chrome Web Store listings; 972,000 declared users (Nov. 3) | The Hacker News |
FAQ
Were people infected?
The coverage read reports no confirmed victims or losses; the numbers are inventories and declared extension users.
What is ClickFix?
A lure that tells the visitor to paste and run a command, here a PowerShell command shown on a fake verification page.
How can I avoid it?
Reach AI services through bookmarks or typed official addresses, and never run commands a web page asks you to paste.
Method and limits
Checked on November 13, 2026. The Hacker News was read in full; Island, Cryptopolitan and Kraven Security were read as search excerpts. Island’s full report was not read directly.
Sources
- The Hacker News: 186 Fake AI Domains Redirect Users to ClickFix Lures, Browser Malware, and Search Hijackers (verified 2026-11-08)
- Island: TypoPrompting: From AI Typos to Malware (verified 2026-11-08)
- Cryptopolitan: 186 fake AI domains funnel ChatGPT and Claude typos into malware (verified 2026-11-08)
- Kraven Security: Triaging the Week 143 (verified 2026-11-08)
Spotted an error in this article? Report it · Our corrections policy