AI Tool Lures Drive 67% of Browser-Based Attacks, Exposing Shadow AI Controls
Source
ClickFix has never been bigger, and AI tools are its most convincing lure yetPush Security / Bleeping Computer
What happened
Push Security's detection data, reported by Bleeping Computer in ClickFix has never been bigger, and AI tools are its most convincing lure yet, shows that ClickFix attacks, a technique that tricks users into running malicious commands by impersonating legitimate software install steps, accounted for 52% of browser-based attack detections through the second quarter of 2026, rising to 67% in August. The primary lures are now fake versions of popular AI tools: attackers clone the install pages of Claude, ChatGPT, and Codex, or abuse AI platform file-sharing features to host malicious payloads. Employees who discover AI tools through search engines rather than approved channels are the target population, as they arrive at look-alike pages before any security review has occurred. Attackers also exploit the fact that many AI tool installs require users to run commands directly in a terminal, which is exactly the step ClickFix hijacks to execute harmful code on the user's machine.
Why it matters
- ·Shadow AI adoption is now a primary attack entry point. Employees searching for unapproved AI tools are being routed to malicious sites, meaning every gap in your AI acceptable-use and procurement approval program translates directly into security exposure.
- ·Existing browser and endpoint controls were not designed for AI-specific lures. Attackers exploit the trust employees place in AI platform branding and the routine nature of terminal-based AI tool installs, bypassing awareness training that assumes employees can spot generic phishing.
- ·This pattern connects to a broader credential theft ecosystem. Stolen credentials from these attacks feed into the proxy networks and supply chain compromises covered in 80,000 Proxy Servers Are Laundering Stolen AI Credentials at Scale, amplifying the downstream risk for any organization whose employees access AI tools outside approved channels.
Stay ahead of stories like this
Get every Global AI governance development like this one, plus the rest of the week's developments. Every Thursday.
Recent issues
- State-linked attackers and criminal operators are actively weaponizing AI tools as entry points into enterprise systems, turning internal platforms and open-source frameworks into live attack vectors.5 Nov
- Three separate incidents this week each created a new compliance action item: verify your AI vendors' data sourcing, demand agent monitoring coverage guarantees, and audit trusted app identities in your cloud environment.29 Oct
Free every Thursday. Unsubscribe anytime.
Governance controls affected
What to do now
- ☐Audit your acceptable-use policy for AI tools to confirm it explicitly prohibits installing AI software found through search engines, and verify employees have been trained on this rule in the past six months.
- ☐Ask your IT or security team whether browser controls are configured to block known AI tool impersonation domains and to alert when users visit AI vendor sites that are not on your approved list.
- ☐Review your approved AI tool list and confirm every entry has a documented install procedure supplied by your IT team, so employees never need to find or follow install instructions from a search result.
- ☐Identify whether any employees routinely run AI tool install commands in a terminal on their work machines, and confirm those workflows go through an IT-managed installation process rather than self-service downloads.
- ☐Check that your vendor due-diligence process (PRC-001) covers AI tools employees are likely to seek out independently, and add a process for employees to request new tools through a governed channel rather than self-installing.
What to watch next
Security teams should monitor whether ClickFix techniques extend to AI platform sharing features, which are already being abused to host malicious payloads. Compliance teams should watch for guidance from regulators on shadow AI and acceptable-use obligations, particularly as enforcement attention on employee-driven AI adoption grows. The connection between browser-based AI credential theft and downstream supply chain compromise means any new disclosures about 80,000 Proxy Servers Are Laundering Stolen AI Credentials at Scale or related proxy abuse should be treated as directly relevant to this threat pattern.