Mr Jones claimed he had warned Microsoft management before but saw no action
Microsoft artificial intelligence engineer Shane Jones raised concerns in a letter on Wednesday. He claimed that the company’s artificial intelligence image generator, Copilot Designer, lacked safeguards to prevent the generation of inappropriate content, such as violent or sexual images. Jones claims he previously warned Microsoft management but saw no action, prompting him to send the letter to the FTC and Microsoft’s board of directors.
“It is very clear within the company that this product is generating harmful images and that these are systemic issues that may be offensive and inappropriate for consumers,” Jones said in the letter posted on LinkedIn. He listed his title as “Principal Software Engineering Manager “.
In response to these accusations, a Microsoft spokesperson denied ignoring security issues. protector the report said. They emphasized the existence of “robust internal reporting channels” to resolve issues related to the generation of AI tools. As of now, Sean Jones has not responded to the spokesperson’s statement.
The central issue raised in the letter concerns Microsoft’s Copilot Designer, an image generation tool powered by OpenAI’s DALL-E 3 system. Its function is to create images based on text prompts.
The incident is part of a broader trend in the field of generative artificial intelligence, which has seen activity surge over the past year. Adding to this rapid development are concerns that AI could be misused to spread disinformation and generate harmful content that promotes misogyny, racism and violence.
“Using only the prompt ‘car crash,’ the co-pilot designer generated an image of a woman kneeling in front of a car wearing only her underwear, including examples of image generation,” Jones said in the letter. “It also generated multiple images Images of women in lingerie sitting on car hoods or walking in front of cars.”
Microsoft has refuted the accusations, saying it has a dedicated team dedicated to evaluating potential security issues in its artificial intelligence tools. Additionally, they claimed to have facilitated a meeting between Jones and the Office of Responsible Artificial Intelligence, indicating a willingness to address his concerns through internal channels.
In a statement to the Guardian, a Microsoft spokesperson said: “We are committed to addressing any employee concerns about our company policies and appreciate the work our employees do to research and test our latest technologies to further improve their security. effort.”
Last year, Microsoft launched its “artificial intelligence companion” Copilot and widely promoted it as a breakthrough way to integrate artificial intelligence tools into business and creative enterprises. Positioned as a user-friendly product for the masses, the company showcased Copilot in a Super Bowl ad last month, emphasizing its accessibility with the slogan “Anyone, Anywhere, Any Device.” Jones believes that it is reckless to portray Copilot Designer as universally safe for use and that Microsoft has neglected to disclose the widely recognized risks associated with the tool.
