Microsoft’s new Azure OpenAI Service terms of service explicitly forbid its integration with any U.S. police department. This includes the use of OpenAI’s text- and speech-analyzing models. Additionally, the terms prohibit law enforcement worldwide from utilizing real-time facial recognition technology on mobile cameras in uncontrolled environments, such as body or dash cameras. These changes were implemented a week after Axon, a company specializing in military and law enforcement technology, unveiled a new product that leverages OpenAI’s GPT-4 generative text model. The updated policy leaves room for Microsoft, as the ban applies solely to U.S. police departments and does not extend to stationary facial recognition systems in controlled environments. However, it aligns with Microsoft’s and OpenAI’s evolving approach to AI-related law enforcement and defense contracts. Microsoft has previously collaborated with the Pentagon on cybersecurity projects, while OpenAI has recently lifted its ban on military AI use. Microsoft’s Azure OpenAI Service was recently incorporated into Azure Government, catering to government agencies, including law enforcement, and has been pledged to undergo additional authorization for DoD mission support. Despite requests for comment, Microsoft and OpenAI have not responded.
Microsoft Restricts Police Use of Generative AI in Azure OpenAI Service
