Beware: Malicious Google Ads Mimic ChatGPT to Phish Users via Windows Run
CCB0x45 · reddit · 2026-08-14
A Reddit user warned of a sophisticated phishing attack they narrowly escaped. While asking ChatGPT a question, they were almost tricked into pasting a malicious command into the Windows Run dialog.
Investigation revealed that this was caused by a malicious Google Ad. Attackers purchase top ad placements on Google search that direct users to a custom GPT or clone site looking exactly like ChatGPT. This fake interface then injects phishing instructions, attempting to trick users into executing dangerous commands. The disguise is so convincing that even experienced users can fall for it if they aren't paying close attention.
More from Safety
- OpenAI Sandbox Escape Sparks Debate: Why Was Internal Proxy Exposed? — SweetDimension7 · 2026-08-14
- Over 800 Fake AI Skills and MCP Servers Found Delivering Malware — HaktanSuren · 2026-08-14
- Cooperative AI Seminar: Solving AI Game Theory Dilemmas with Safe Pareto Improvements — xuanalogue · 2026-08-14
- Stanford HAI: Science Needs Truly Open Source AI, Not Just Open Weights — StanfordHAI · 2026-08-14
- METR and Redwood Urged to Disclose OpenAI Safety Audit Terms — DKokotajlo · 2026-08-14
- Should AI Developers Refuse to Work with Oppressive Governments? — deanwball · 2026-08-14