As part of its new Responsible AI Standard rules, Microsoft plans to remove Azure's facial analysis tools, which can guess gender, age, and emotional state
The technology giant will stop offering automated tools that predict a person's gender, age and emotional state and will restrict the use of its facial recognition tool.
New York TimesKashmir Hill
Context & Ripple Effects
Microsoft had already stepped back from third-party facial-recognition investment through its exit from firms including AnyVision, while its AI ethics leadership said the company had rejected some facial-recognition client requests. The Responsible AI Standard turns that case-by-case posture into an Azure product and access policy.
The move also aligns with the regulatory pressure around automatic gender-recognition systems, whose underlying assumptions were challenged by NGOs as the EU weighed AI rules.
First-order effects
Azure customers lose access to tools that infer gender, age, and emotional state, and face tighter conditions on Microsoft's remaining facial-recognition offering.
Microsoft shifts responsibility for limiting these uses from individual sales decisions to the Azure service itself under its Responsible AI Standard.
Second-order effects
Organizations using Azure for demographic or emotion inference must redesign those workflows or seek alternatives, while other cloud vendors face a clearer benchmark for restricting sensitive facial-analysis features.
Microsoft's prior rejection of some facial-recognition requests becomes a more durable commercial constraint: customers cannot treat Azure availability as approval for every proposed deployment.
Third-order effects
If major AI providers embed use restrictions directly in product availability and access terms, governance shifts from voluntary principles toward provider-enforced controls at the cloud layer.
The dispute over automated gender recognition points to a broader split between identity- or emotion-inference products and tools providers are willing to support, particularly as regulation develops.
The trend: Cloud AI providers are increasingly making sensitive-use governance enforceable through product removals and access restrictions rather than relying only on stated principles.
Microsoft will no longer offer facial analysis tools that predict age, gender and emotion as part of a broader push for more responsible AI. It will also restrict its facial recognition tool to those who apply and explain how it will be used. https://www.nytimes.com/...
wonder whether GPT-3 and other LLM's like LaMDA will similarly eventually be dialed back because of their tendency towards generating misinformation https://twitter.com/...
Very interesting reading, and glad to see this decision to retire the scientifically and ethically dubious features of AI facial analysis that claim to “infer emotional states and identity attributes” from faces https://blogs.microsoft.com/ ...
Today, @Microsoft is sharing our Responsible AI Standard to help guide the development of trustworthy #AI. We hope our AI journey will contribute to responsible AI policy discussions and industry practices. Read more: https://blogs.microsoft.com/ ...
Good development that also illustrates the limits of the corporate ethical AI. Microsoft will halt one type of products but still partners with, supports, and offers critical infrastructure for other surveillance products & systems used across the world https://www.nytimes.com/..…
Prediction: other (responsible) big tech companies will get out of this space. Recognising identity, emotions, gender, age ... using automated tools will all soon be considered caustic https://www.nytimes.com/...
A lot of thought has gone into this and there is still a lot more to learn. Today, Microsoft externalized our Responsible AI Standard. Honored to have contributed to it, and proud to work for a company that holds AI products to a standard. https://blogs.microsoft.com/ ...
Today, we released our updated Responsible AI Standard, an important step toward developing better, more trustworthy AI. It's a living document that will evolve to address new research, technologies, laws, & learnings from within & outside the company. 👇 https://blogs.microsoft.c…
Microsoft stops offering automated inferences that predict gender, age, emotional state & will restrict use of its facial analysis tool. Some background on our programs & efforts aimed at responsible development & fielding of #AI https://www.nytimes.com/... @PartnershipAI @MSFTIs…
A testament to the power of data justice-oriented research and advocacy & a reminder to keep it up! Facial analysis tools, among other spurious technologies, are NOT inevitable! https://www.nytimes.com/...
“Ms. Crampton said Microsoft had considered whether to start making its software available to the police in states with laws on the books but had decided, for now, not to do so. She said that could change as the legal landscape changed.” https://www.nytimes.com/...
Microsoft is advancing AI with ethical principles that put people first. Our Responsible AI Standard takes a thoughtful approach, moving from principles to practices and collaborating with industry, academia and regulators for meaningful progress. https://blogs.microsoft.com/ ...
💡We recognize the Responsible AI Standard is just one of many steps in our Responsible AI journey, and we're committed to open, honest and transparent in our efforts to make meaningful progress. Learn more here: https://blogs.microsoft.com/ ... (4/5)
Interesting development! Did Microsoft want to do the right thing or is it partly for self-serving reasons? My take: it doesn't matter. Either way, it shows that research, journalism, and advocacy—both internal and external—can change companies' behavior. https://www.nytimes.com/…
Kudos to @Microsoft! “The technology giant will stop offering automated tools that predict a person's gender, age and emotional state and will restrict the use of its facial recognition tool.” See also: https://papers.ssrn.com/... https://twitter.com/...
“This is a critical norm-setting period for A.I.” - Natasha Crampton, our Chief Responsible AI Officer, speaking with the @NYTimes about @Microsoft's updated Responsible AI Standard paving the way for the tech industry. https://www.nytimes.com/...
Story behind the story for my fellow Responsible AI professionals is Microsoft's release today of its v2 Responsible AI framework incl. assessment templates, a github repository, and more. Kudos to the team over there for their trailblazing work. https://blogs.microsoft.com/ ... …
Computational infrastructures at work to normalize biometrics for authentication (interesting that it is only GApple + Microsoft and they left out Amazon) and tracking. Facial recognition or fingerprint as the basis of access to resources should be very concerning! https://twitte…