No question about This: Nadella’s Microsoft is a triumph. Finally, in the 2020s, Microsoft itself has focused on the most advanced technology since the PC. And while revenue from AI products hasn’t begun to offset Microsoft’s heavy investment, it has the confidence and resources to wait until the products improve and become useful to users.
But can Microsoft really escape the hubris that set it up so long ago? Consider what happened in May of this year with a product called Recall.
The feature was supposed to be a manifestation of Microsoft’s integration of AI into its hardware, software and infrastructure. The idea was to give users something like a personal version of the Internet Archive. Remembering will permanently capture everything that happens on your machine: what you read, what you write, the photos and videos you view, the sites you visit. Just describe to your machine what you are looking for: What were the rug patterns I was considering for my living room? Where is that report on Amazon’s environment? When did I go to Paris? Those moments will pop up like magic, as if you have a homunculus who knows everything about you. It sounds scary – like the Big Brother of the ship – but Microsoft insists that users can 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 messed around with your personal information, no matter how sensitive, without asking for permission. While Microsoft has stressed that only the user can access Recall, security researchers have found “flaws that could let you fly the plane,” as one tester put it.
“Within about 48 hours, we were like, ‘Wow, this is incredibly exciting!’ For people expressing some concerns,” says Brad Smith. While the press piled in, Smith was on a plane to meet Nadella in Washington, DC. By the time he landed, he thought it would make sense to make the recall work only if users opted in. Nadella agreed. Meanwhile, in Redmond, senior Microsoft executives piled into meeting rooms to see how they could get the product back. Fortunately, since the feature hadn’t arrived yet, they didn’t have to recall the recall. They postponed the launch. And they’ll add security features, like “just-in-time” encryption.
“People pointed out some obvious things to us that we should have caught,” Nadella says. But even his responsible AI team missed him. A measure of “knowing it all” led to a product announcement that fell short, indicating that, even when led by a purported sympathizer, Microsoft still held its former position. Retains many flaws. Only now, it’s a $3 trillion company with access to leading AI operations products.
“You can think of it in one of two ways,” says Brad Smith. One is, ‘Gee, I wish we’d thought of that earlier.’ Hindsight is a great thing. Or two, ‘Hey, it’s cool that we’re using this to make this change. Let us explain why.’ It was truly a learning moment for the entire company.
That’s right. After 50 years, though, it’s a lesson 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. mail@wired.com.
