Microsoft employee Shane Jones has raised concerns about the safety of one of the company's generative AI tools, Copilot Designer, which has the ability to produce potentially inappropriate images. In a letter to the Federal Trade Commission and Microsoft's board of directors, Jones expressed that the AI tool generated images containing sex, violence, underage drinking, drug use, political bias, misuse of corporate trademarks, and conspiracy theories.
Jones emphasized the importance of educating the public, particularly parents and teachers, about the risks associated with using Copilot Designer, especially for school projects or educational purposes involving children. He noted that the AI image generator could add harmful content to images created from seemingly benign prompts, citing examples such as inappropriate depictions related to prompts like "car accident," "pro-choice," and "teenagers 420 party."
Despite Jones' efforts to urge Microsoft to address the issues and implement better safeguards, the company reportedly declined his recommendations. Microsoft has stated that it is committed to addressing any concerns raised by employees and appreciates efforts to test and enhance the safety of its technology.
This isn't the first time Jones has voiced concerns about AI image generators. Prior to his letter to the FTC, he reportedly posted an open letter on LinkedIn to OpenAI about removing DALL-E, the model powering Copilot Designer, from public use. In late January, Jones sent a letter to US senators regarding public safety risks associated with AI image generators and alleged efforts by Microsoft to silence his concerns.
In light of Google pausing access to its image generation feature on Gemini due to similar concerns, Jones praised Google's swift action and called on Microsoft to lead in addressing issues related to AI image generators. The response from Microsoft and the FTC to Jones' letter remains pending, with both parties yet to comment on the matter.