Microsoft will stop using facial recognition technology with which the company says it can guess people’s emotions, among other things. According to Microsoft, the artificial intelligence raised questions about privacy protection and the risk of discrimination and other forms of abuse.
It concerns facial recognition in the Azure Face service, which Microsoft sells to companies. Not only does the technology recognize which emotions a person shows, but also gender, age, hair and makeup.
Microsoft worked with researchers to list the pros and cons of the technology. “Specifically in the case of emotion recognition, this raised many questions about privacy and the lack of a clear definition of the word emotion,” the company writes. “The same goes for the inability to relate facial expressions to the emotional state of people in different situations and regions.”
To limit the risk of misuse, the service will be discontinued. New users can no longer purchase the service, existing customers will have access until June 30, 2023.
Criticism of facial recognition systems for some time
Experts have long been negative about the consequences of programs that can read faces. For example, they could lead to discrimination among employers when hiring new people. In addition, lawmakers in the US and Europe are discussing the legality of the technology.
IBM stopped developing facial recognition two years ago. According to the company, the systems are often biased and can be used for mass surveillance and ethnic profiling.
Although Microsoft will stop certain parts of its facial recognition, some will remain available. It concerns Seeing AI, suitable for people with a visual impairment. Image recognition is used in the app to describe the environment.