Microsoft AI engineer Shane Jones has publicly raised concerns about potential safety issues with OpenAI's DALL-E 3, specifically its ability to generate violent or explicit content.

Shane Jones claims that Microsoft's legal team prevented him from speaking out on these vulnerabilities and has escalated the issue to the FTC.

Experiments by CNBC have confirmed that DALL-E 3 can be manipulated into producing objectionable content with seemingly innocuous inputs.

Microsoft reportedly faces over 1,000 product feedback complaints daily and struggles with resource allocation to address and rectify these concerns.

Despite OpenAI and Microsoft asserting that safety measures are in place, Jones warns there is inadequate infrastructure for reporting and mitigating the spread of harmful content created by DALL-E 3.

The incident echoes a recent problem with Google's Gemini chatbot, which produced historically inaccurate images, prompting Google to suspend its image generation service pending a solution.