Microsoft is phasing out public access to a number of AI-driven facial analysis tools, including one that seeks to identify a subject’s emotion from videos and images.
These “emotion recognition” tools have been criticized by experts. They say that not only are facial expressions believed to be universal differing between different populations, but it is unscientific to equate external manifestations of emotion with internal feelings.
“Businesses can say whatever they want, but the data is clear,” Lisa Feldman Barrett, a psychology professor at Northeastern University who conducted a review on the topic of AI-driven emotion recognition, told The Verge in 2019. . “They can detect a face to face, but that’s not the same as detecting anger.”
“They can detect a face to face, but that’s not the same as detecting anger.”
The decision is part of a broader review of Microsoft’s AI ethics policies. The company’s updated responsible AI standards (first described in 2019) emphasize responsibility for finding out who uses its services and greater human oversight of where these tools are applied.
In practical terms, this means that Microsoft will restrict access to some features of its facial recognition services (known as Azure Face) and completely eliminate others. Users will be asked to use Azure Face for facial identification, for example, telling Microsoft exactly how and where to deploy their systems. Some use cases with less harmful potential (such as automatically blurred faces in images and videos) will remain open access.
In addition to removing public access to its emotion recognition tool, Microsoft also removes Azure Face’s ability to identify “attributes such as gender, age, smile, facial hair, hair, and the makeup “.
Experts inside and outside the company have highlighted the lack of scientific consensus on the definition of ’emotions’, the challenges of how inferences are generalized between use cases, regions and demographics, and the highest concerns of privacy around this type of capacity, “he wrote. Microsoft’s head of AI, Natasha Crampton, in a blog post announcing the news.
Microsoft says it will stop offering these features to new customers starting today, June 21, while existing customers will be denied access on June 30, 2023.
Microsoft’s Seeing AI application will still offer emotion recognition
However, as long as Microsoft withdraws public access to these features, it will continue to use them in at least one of its own products: an application called Seeing AI that uses artificial vision to describe the world of the visually impaired.
In a blog post, Microsoft Group’s chief product officer for Azure AI, Sarah Bird, said that tools such as “emotion recognition” can be valuable when used for a set of scenarios. controlled accessibility “. It is unclear whether these tools will be used in other Microsoft products.
Microsoft is also introducing restrictions similar to its custom neural voice feature, which allows customers to create AI voices based on real-person recordings (sometimes known as audio deepfake).
The tool “has exciting potential in education, accessibility and entertainment,” writes Bird, but notes that “it’s also easy to imagine how it could be used to improperly impersonate speakers and mislead listeners.” Microsoft says that in the future it will limit access to the feature to “managed clients and partners” and “ensure the active participation of the speaker when a synthetic voice is created.”