OpenAI works to stop ChatGPT generating 'sex crime scene' images

Researchers at British AI security startup Mindgard have discovered that the latest public version of OpenAI's ChatGPT can be manipulated to generate sexualized and violent images, including what they described as 'sex crime scene' depictions.

OpenAI stated that it took action to prevent the chatbot from responding with such images. The company did not disclose specific technical details of the fix but emphasized its commitment to safety. This is not the first time ChatGPT has been shown to produce inappropriate content; similar vulnerabilities have been reported previously, often leading to temporary restrictions or updates.

The incident highlights ongoing challenges in ensuring that large language models with image generation capabilities adhere to content policies. It also underscores the role of external researchers in identifying flaws before they can be exploited at scale. OpenAI has a history of working with such researchers to address reported issues.