No questions about this: Nadella’s Microsoft is a triumph. Finally, in the 2020s, Microsoft focused on the most innovative technology since the PC itself. And while revenue from AI products hasn’t begun to offset Microsoft’s huge investments, the company has the confidence (and resources) to wait until products improve and users find them useful.
But can Microsoft really avoid the hubris that marked it so far before? Consider what happened in May of this year with a product called Recall.
This feature was meant to embody Microsoft’s integration of AI into its hardware, software, and infrastructure. The idea was to give users a sort of personal version of Internet Archive. Recall would constantly capture everything that happens on your machine: what you read, what you write, the photos and videos you watch, the sites you visit. Simply describe to your machine what you’re looking for: What were the rug samples I was considering for my living room? Where is this report on the ecology of the Amazon? When did I go to Paris? These moments would appear as if by magic, as if you had a homunculus who knew everything about you. It sounds scary – a bit like an embedded Big Brother – but Microsoft insisted users could feel safe. Everything stays on your computer!
Almost immediately, critics called it a privacy nightmare. For one thing, they noted, Recall worked by default and gobbled up your personal information, no matter how sensitive, without asking permission. While Microsoft emphasized that only the user could access Recall, security researchers found “gaps you could drive a plane into“, as one tester said.
“In about 48 hours, we went from, ‘Wow, this is extraordinarily exciting!’ to people expressing certain reservations,” says Brad Smith. As the press piled in, Smith was on a plane to meet Nadella in Washington, DC. By the time he landed, he thought it would be prudent to run Recall only if users consented to it; Nadella agreed. Meanwhile, in Redmond, Microsoft’s top executives crowded into meeting rooms to see how they could shrink the product. Luckily, since the feature wasn’t available yet, they didn’t need to bring Recall back. They postponed the launch. And they would add security features, like “just-in-time” encryption.
“People pointed out some obvious things we should have done,” Nadella says. But his own responsible AI team also missed them. A certain amount of know-it-all led to the announcement of a product that fell short, indicating that, even led by a supposedly empathetic person, Microsoft still retains too many of its earlier character flaws . Only now it’s a $3 trillion company with secure access to cutting-edge AI operation products.
“You can think of it two ways,” says Brad Smith. “The first is, ‘Well, I wish we’d thought about this first.’ » Hindsight is a good thing. Or two: Hey, it’s great that we’re using this to make this change; Let’s be explicit about why. » It was truly a learning moment for the entire company.
It’s very good. After 50 years, this is a lesson that Microsoft – and Nadella – should have learned long ago.
Getty Images (timeline)
Let us know what you think about this article. Send a letter to the editor to mail@wired.com.