Microsoft limits Azure AI facial recognition use for police forces

AH Microsoft Logo (1)

It seems that Microsoft does not want law enforcement to use Azure OpenAI for real-time facial recognition of individuals. The Redmond company recently modified its code of conduct regarding the use of its AI services by the police.

Law enforcement agencies have also been adapting to AI-based technologies for some time now. Additionally, suppliers of military and police technology, such as Axon. For example, the latter recently announced that it would draw on the capabilities of the GPT-4 model to facilitate transcriptions of audio captured by mobile cameras. This can be quite useful for speeding things up in police reports.

However, the latest modifications to the Azure AI code of conduct are aimed directly at mobile cameras. As spotted by The Register, Microsoft strictly banned the use of Azure AI for real-time facial recognition using devices such as on-body police cams and dashcams. Microsoft extended the limitations to police forces around the world. Previously, the ban was active only for the US state and local police.

Microsoft doesn’t want police to use Azure AI for real-time facial recognition

Regarding the ban, the new code of conduct specifically mentions some key phrases. For example, “real-time facial recognition technology on mobile cameras” and “in the wild environments.” This suggests that police forces could still use Azure AI for facial recognition on content captured by security cameras. Also on recordings after situations. But there is no confirmation on this yet.

The exact reason behind Microsoft’s decision is not known. The company did not respond to questions from The Register about it. However, it could go in line with one of the company’s core policies regarding the use and implementation of AI-based technologies. More specifically, among its policies it says that it cannot be “used to infer sensitive attributes of people, such as gender, race or specific age, from images or videos of them.”

Additionally, the use of AI has shown some problems at times when it comes to identification by attributes. There have been cases where, in some circumstances, it has shown biases of some kind. Therefore, its use is not totally foolproof in this area, at least today.

The post Microsoft limits Azure AI facial recognition use for police forces appeared first on Android Headlines.