Microsoft is abandoning its scientifically and ethically questionable emotion recognition technology. At least for now. In a big win for privacy advocates alarming about poorly tested and invasive biometric technologies, Microsoft has announced plans to remove “emotion recognition” detection systems from Azure Facial recognition services. The company will also gradually abandon the opportunities that artificial intelligence tries to use to determine personality traits such as gender and age.

Microsoft’s decision to slow down the controversial technology comes amid a major overhaul of its ethics policies. Natasha Crampton, Microsoft’s chief artificial intelligence officer, said the company’s decision came in response to experts who said there was a lack of consensus on “emotions” and fears of overgeneralizing how AI systems might interpret those emotions.

“We worked with internal and external researchers to understand the limitations and potential benefits of this technology and find trade-offs,” said Sarah Byrd, product manager at Azure AI Principal Group. “API access to opportunities involving hidden attributes also opens up a wide variety of ways to misuse them, including stereotyping, discrimination, or unfair denial of service,” Byrd added.

Byrd said the company will abandon the general-purpose system in the Azure Face API, which aims to measure these features to mitigate risks. Starting Tuesday, new Azure customers will no longer have access to the detection system, but existing customers will be able to terminate it by 2023. While Microsoft has said its API will no longer be generally available, Byrd said the company may explore the technology in some limited use cases, including as a tool to support people with disabilities.

“Microsoft recognizes that these capabilities can be useful when used for a number of controlled access scenarios,” Byrd said.

The course is being adjusted to align Microsoft’s policy with the new 27-page Standard for Responsible Artificial Intelligence, which has been in development for a year. Among other recommendations, the standard calls on Microsoft to ensure proper data management in its products, provide informed human oversight and control, and “provide effective solutions to targeted problems.” Source