ChatGPT Search Tool Vulnerable to Manipulation and Deception

OpenAI’s ChatGPT search tool, designed to enhance online queries, has been found vulnerable to manipulation via hidden content, according to a Guardian investigation. Techniques like “prompt injection” can influence the AI’s responses, allowing it to promote misleading information or even deliver malicious code from websites. Security researchers warn that such flaws could lead to misuse, such as biased product reviews or phishing scams, and emphasise the need for safeguards.

Read more about this issue on The Guardian here.