AI BriefingChatGPT Found Recommending Fraudulent Fake Shopping Sites
Key point
ChatGPT has raised security concerns by recommending fake websites of bankrupt brands due to training data contamination.
Details
Cases have been reported of ChatGPT recommending fraudulent online shopping sites and fake websites to users. Investigation revealed that the system is directing users to cloned websites of brands that have already gone bankrupt.
Experts have raised the possibility of data poisoning, where malicious content is included in the model's training data, as a cause of this phenomenon.
This security vulnerability is considered a particularly serious issue at a time when big tech companies such as Amazon and Google are accelerating the introduction of AI Agents that make purchases on behalf of users. Authorities are recommending that users always double-check the information recommended by chatbots.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Fake Customer Service Pages Induce ChatGPT and Gemini to Display Scam Numbers
Australian PM Reports OpenAI Agent Hacked Medicare Portal in Believed First Known AI-Driven Government BreachRootless Linux Sandbox 'Drop' Released for Isolating AI Agents and Third-Party Programs