No question is: Nadella’s Microsoft is a triumph. Finally, in the 1920s, Microsoft focused on the most groundbreaking technology since the PC itself. And while AI product revenues haven’t yet begun to offset Microsoft’s massive investments, Microsoft has the confidence — and resources — to wait until the products are improved and users find them useful.
But can Microsoft really avoid the hubris that caused it this far back? Consider what happened in May this year with a product called Recall.
This feature was intended to embody Microsoft’s integration of artificial intelligence with hardware, software, and infrastructure. The idea was to provide users with a sort of personal version of the Internet Archive. Recall will constantly record everything that happens on your computer: what you read, what you write, photos and videos you view, websites you visit. Just describe to your machine what you’re looking for: What rug samples was I considering for my living room? Where is this report on the ecology of the Amazon? When did I go to Paris? These moments appeared like magic, as if you had a homunculus that knew everything about you. It sounds scary – a bit like on-board Huge Brother – but Microsoft insisted that users should feel protected. Everything stays on your computer!
Almost immediately, critics labeled it a privacy nightmare. First, they noticed that Recall worked by default and absorbed your personal data, no matter how sensitive, without asking for permission. Although Microsoft emphasized that only the user could access the Recall feature, security researchers discovered “gaps through which an airplane could fly” as one tester put it.
“Within about 48 hours, we were like, ‘Wow, this is incredibly exciting!’ people expressing some reservations,” says Brad Smith. While the press was piling on, Smith was on a plane to meet Nadella in Washington. Before he landed, he concluded that it would be wise for the restore function to only work if users had consented to it; Nadella agreed. Meanwhile, in Redmond, senior Microsoft executives gathered in conference rooms to see how they could narrow the scope of the product. Fortunately, since the feature wasn’t shipped yet, they didn’t have to deprecate Recall. They postponed the premiere. They would also add security features like just-in-time encryption.
“People pointed out some obvious things that we should have done that we should have caught,” says Nadella. But his own responsible AI team missed them too. A certain measure of “know-it-all” led to a product announcement that fell low, indicating that Microsoft, even led by a supposed empath, still retained too many of its previous character flaws. Only now it’s a $3 trillion company with closed access to products with a cutting-edge AI operation.
“There are two ways to think about it,” says Brad Smith. “One of them is, ‘Oh man, I wish we had thought of that sooner.’ Looking back is a wonderful thing. Or two: “Hey, it’s good that we’re using this to make this change – let’s be clear about why.” It was truly a learning moment for the entire company.”
All right. But 50 years later, it’s a lesson Microsoft — and Nadella — should have learned long ago.
Getty Images (timeline)
Let us know what you think about this article. Write to the editor at: mail@wired.com.