A Guardian investigation finds that OpenAI's ChatGPT search tool can be manipulated, leading to false or harmful search results due to hidden text on webpages.

OpenAI's latest offering, the ChatGPT search tool, has come under scrutiny as recent tests conducted by The Guardian reveal significant vulnerabilities. This AI-driven search tool, designed to provide streamlined and efficient results, may inadvertently expose users to deceptive content due to its reliance on hidden text found on websites. As OpenAI encourages a wider user base to adopt this tool as their default search engine, the implications of these findings raise concerns about security and reliability in the increasingly complex landscape of AI-powered technology.

The investigation pointed out that the ChatGPT search tool is susceptible to manipulation where malicious actors can embed hidden content within web pages, effectively tricking the AI into retrieving harmful information. OpenAI has made this tool available to paying customers, prompting calls for enhanced scrutiny and potentially necessary adjustments to minimize risks. With the rise in cyber threats, addressing these vulnerabilities is crucial for establishing trust among users and ensuring a safer online search experience.