Whistleblower Microsoft Engineer Exposes Harmful AI Image-Generator

Microsoft Engineer Raises Concerns About Offensive Imagery Generated by AI Tool

A Microsoft engineer, Shane Jones, has raised alarms about offensive and harmful imagery generated by the company’s artificial intelligence image-generator tool, Copilot Designer. Jones sent letters to U.S. regulators and Microsoft’s board of directors urging them to take action to address the safety concerns.

  • Whistleblower: Jones considers himself a whistleblower and recently met with U.S. Senate staffers to share his concerns.
  • FTC Confirmation: The Federal Trade Commission confirmed receiving Jones’s letter but declined further comment.
  • Microsoft’s Response: Microsoft stated that it is committed to addressing employee concerns and appreciates Jones’ efforts in studying and testing the technology.

Jones expressed his safety concerns about Copilot Designer, citing instances where the tool generated inappropriate and harmful content despite benign user prompts. He highlighted issues such as sexually objectified imagery, violence, political bias, and more.

Jones’s Efforts and Public Disclosure

Jones has been persistent in addressing the safety concerns surrounding Copilot Designer. He first raised his findings with OpenAI, Microsoft’s business partner, before escalating the matter to U.S. regulators and Microsoft’s board of directors. Additionally, he brought his concerns to the state attorney general in Washington.

Despite Microsoft’s recommendation to use internal reporting channels, Jones took his concerns public, posting a letter on LinkedIn that was later removed at the request of Microsoft’s legal team.

Concerns About AI Image-Generators

The emergence of advanced AI image-generators like DALL-E 2 and ChatGPT has sparked public fascination but also raised concerns about potential dangers. Without effective safeguards, users can easily generate harmful “deepfake” images that falsely depict real people in compromising situations.

Google’s suspension of its Gemini chatbot’s ability to generate images of people due to concerns about racial depictions underscores the need for responsible AI development and implementation.

As the debate over AI ethics and safety continues, Jones’s advocacy for addressing harmful content in AI-generated imagery serves as a reminder of the importance of responsible technology development.

Subscribe to the Eye on AI newsletter to stay informed about how AI is shaping the future of business. Sign up for free.

About The Author

Scroll to Top