For years, activists and academics have been raising concerns that facial analysis software that claims to be able to identify a person’s age, gender and emotional state can be biased, unreliable or invasive — and shouldn’t be sold.

Acknowledging some of those criticisms, Microsoft said on Tuesday that it planned to remove those features from its artificial intelligence service for detecting, analyzing and recognizing faces. They will stop being available to new users this week, and will be phased out for existing users within the year.

Microsoft has announced it will stop selling technology that guesses someone's emotion based on a facial image and would no longer provide unfettered access to facial recognition technology.

The actions reflect efforts by leading cloud providers to rein in sensitive technologies on their own as lawmakers in the United States and Europe continue to weigh comprehensive legal limits.

Since at least last year, Microsoft has been reviewing whether emotion recognition systems are rooted in science.

"These efforts raised important questions about privacy, the lack of consensus on a definition of 'emotions' and the inability to generalise the linkage between facial expression and emotional state across use cases, regions, and demographics," Sarah Bird, principal group product manager at Microsoft's Azure AI unit, said in a blog post.

Existing customers will have one year before losing access to artificial intelligence tools that purport to infer emotion.

In addition Microsoft is also retiring Azure Face’s ability to identify "attributes such as gender, age, smile, facial hair, hair, and makeup", according to the blog post.

Similar restrictions are also being added to the company's Custom Neural Voice feature, with which customers are able to create AI voices based on recordings.

While the tool had "exciting potential in education, accessibility, and entertainment… it is also easy to imagine how it could be used to inappropriately impersonate speakers and deceive listeners," Bird wrote.

The functionality will be limited to "managed customers and partners" with the "active participation" of the speaker necessary when creating a synthetic voice in the future.


Alphabet's Google Cloud last year embarked on a similar evaluation of AI tools, first reported by Reuters.

Google blocked 13 planned emotions from its tool for reading emotion and placed under review four existing ones, such as joy and sorrow.

It was weighing a new system that would describe movements such as frowning and smiling, without seeking to attach them to an emotion.

Google did not immediately respond to a request for comment.

Microsoft also said customers now must obtain approval to use its facial recognition services, which can enable people to log into websites or open locked doors through a face scan.

The company called on clients to avoid situations that infringe on privacy or in which the technology might struggle, such as identifying minors, but did not explicitly ban those uses.