A secnior Microsoft employee has called for an investigation by the US government into the safety of one of the company's generative AI products. Microsoft's board of directors and the US Federal Trade Commission have received a letter from Shane Jones, a principal software engineering manager, concerning Copilot Designer, the company's text-to-image converter that was introduced in March 2023.

In his letter, Jones voiced concern about Copilot Designer's ability to create potentially offensive visuals, such as those showing violence, sex, drug usage, underage drinking, political prejudice, and conspiracy theories. He emphasized how crucial it is to inform the public—in particular, parents and educators—about the dangers of utilizing this kind of technology, particularly in learning environments.

Jones disclosed that over the last three months, he had repeatedly attempted to bring the matter up with Microsoft internally, but the corporation had refused to take any action or withdraw Copilot Designer from public usage. He recommended that the product's rating be changed on the Android app store and that disclosures be added, however none of these suggestions was carried out.

In response to Jones's worries, Microsoft said that it values efforts to improve the safety of its technology and is dedicated to handling any employee complaints in accordance with its standards.

Jones has already voiced worries over the safety of AI. He had openly encouraged OpenAI to take DALL-E, the model that powers Copilot Designer, out of public usage months before he wrote the letter to the FTC. Jones persisted in voicing his worries, including contacting US lawmakers over the safety issues associated with AI, in spite of pressure from Microsoft's legal team to take back his remarks.

The event occurs as AI technologies in the IT industry come under increasing scrutiny. In response to concerns of historically incorrect photos depicting race, Google has suspended access to its image generating capability on Gemini, its rival to OpenAI's ChatGPT. The CEO of Google's AI branch, DeepMind, Demis Hassabis, promised that the functionality will be brought back after the issues were resolved.