From Medianama.: 2024-05-03 07:25:26

Microsoft has updated its code of conduct to restrict US police departments from using Azure OpenAI Service integration for facial recognition. The service can no longer be used by law enforcement globally for real-time monitoring on mobile cameras. This comes after Microsoft’s previous stance against selling facial recognition technology to US police.

Facial recognition technology has faced scrutiny due to inaccuracies, with studies showing an 80 percent accuracy rate in matching faces. AI models can also exhibit biases, such as considering someone with drooping eyelids as asleep at the wheel. Microsoft’s updated code of conduct includes restrictions on using AI in subliminal or manipulative ways beyond a person’s consciousness.

The addition to the code of conduct prohibits exploiting vulnerabilities, categorizing people based on biometric data for characteristics like race or political opinions. The move reflects a wider debate on the ethical use of AI and facial recognition technology. Overall, concerns remain over the potential misuse and biases inherent in AI-powered systems.



Read more at Medianama.: Microsoft restricts US police use of Azure OpenAI facial recognition