Microsoft Tightens Grip On Azure’s AI For Facial Recognition

In a discreetly pushed policy update, Microsoft’s Azure OpenAI Service has tightened its hold on the use of Azure’s AI for facial recognition by police forces across the world. First spotted by The Register, the revised code of conduct prohibits real-time facial recognition on mobile cameras by any law enforcement in the world. The revised code also includes on-body police cams and dashcams.

Previously, these limitations mainly focused on US state and local police, however, the latest restrictions from the company extend these limitations to law enforcement agencies around the world.

The policy reads that “real-time facial recognition technology on mobile cameras” and “in the wild environments.” It suggests that the company hasn’t banned agencies from outside of the US from using its cloud service for facial recognition on fixed cams. It also has no mention of post-event processing at a police station, but this seems to be a mere speculation.

Also Read: Apple To Give Siri For iOS 18 Major Upgrade Via Apple’s Ajax LLM

Microsoft Doesn’t Want To Interfere With People's Privacy

The Register also Microsoft for a comment on the matter, but the company declined to explain why it tweaked its policy. However, Axon, a military and law enforcement tech firm announced that it would leverage GPT-4 to help transcribe audio from body cams for police reports.

Despite AI’s potential to transform industries and boost efficiency, it also has a lot of challenges. For instance, AI has been proven to show bias in some scenarios. Microsoft’s policy reads that Azure OpenAI Service must not “be used to infer people’s sensitive attributes such as gender, race, or specific age from images or videos of them.”

The policy is applicable to all OpenAI models, including GPT and DALL-E tools. As the debate about the ethical use of AI continues, Microsoft’s new guidelines can be seen as a notable step in regulating the application of AI-based facial recognition tech - a testament to Microsoft’s claimed commitment to being responsible for AI.

Also Read: OpenAI Reportedly Working On Search Engine To Rival Google

Microsoft Developing New AI Model To Compete With Google

Microsoft seems to have a new in-house AI language model in the works. The new model is said to be big enough to brush shoulders with the model developed by Alphabet’s Google and OpenAI. The new model codenamed MAI-1, is overseen by Mustafa Suleyman, who is the co-founder of Google Deepmind and ex-CEO of Inflection AI startup.  

It’s unclear what Microsoft plans to do with the new model, but it will depend on the performance of the model. The company could announce the new model during its Build developer conference later this month. MAI-1 will be “far larger” compared to its previous smaller, open-source models that the company has trained before.