When it comes to artificial intelligence, Microsoft is currently stealing the limelight. On Tuesday, the company released an update to its Bing search engine that incorporates technology from ChatGPT maker OpenAI, which it’s been cozying up to for years.
Google, long the dominant player in search—with Bing a distant second—has been caught off guard by the speed of Microsoft’s A.I. search rollout, and the Windows maker plans to eventually add OpenAI technology to many other products as well. As Microsoft CEO Satya Nadella told The Verge this week, Google is the “800-pound gorilla” but he hopes Microsoft’s A.I. moves will make it “come out and show that they can dance.”
Fortune reached out to Google for comment but did not receive an immediate reply.
But it may have been an abject failure that helped Microsoft get to this point. In 2016, the company launched an A.I. chatbot via Twitter that it had to take down within a day. Called Tay, the bot mimicked the deliberately offensive behavior of other Twitter users and spewed out inflammatory tweets of its own.
“One of the interesting things that we learned from it wasn’t like, ‘Oh my god, like we should never do this again.’ It’s like, ‘Oh my god, we need to really get serious about responsible A.I. safety and ethics,’” Microsoft CTO Kevin Scott said in an interview with the Hard Fork podcast released today.
“We want to build technology such that it gets the best of humanity, not the worst,” Nadella said at a Microsoft event in March 2016, shortly after the Tay debacle.
Scott added, “It was one of the catalyzing events because we’re still very, very serious about A.I. and had high conviction that it was going to be one of the most important—if not the most important—technology that any of us have ever seen, that we had to go solve all of these problems to avoid making similar sort of mistakes again.”
That set Microsoft up to move quickly on partnering with OpenAI. In 2019, it invested $1 billion in the venture, and then quietly sank another $2 billion into it in 2021. Last month, it announced it would invest billions more into OpenAI.
That initial interest in A.I. has led to the new and improved Bing, which now offers alongside traditional search results the kind of human-sounding answers that have wowed ChatGPT users over the past few months. There’s a waitlist for trying out this version of it, but Microsoft plans to make it available to millions of users in the coming weeks. (You can skip ahead in line by getting the Bing app, which has seen a sudden spike in downloads.)
Google this week announced a ChatGPT rival called Bard, but it hasn’t yet incorporated it into the search results users see. It offered a live-stream presentation, but that was marred by Bard coming up with a factual error. The company lost $100 million in market value after the demo, furthering the impression it’s losing ground to Microsoft.
Scott acknowledged to Hard Fork that Microsoft will still likely make mistakes with the OpenAI-powered Bing. But, he added, since the Tay disaster the company has tried to articulate A.I. principles, not just for marketing purposes but to incorporate them into the work it’s doing.
“And so in that sense, when this new breakthrough came from OpenAI, we didn’t have as much of that resistance as you would imagine. They knew what muscle they were going to go exercise to ensure that we could launch it with confidence.”
Learn how to navigate and strengthen trust in your business with The Trust Factor, a weekly newsletter examining what leaders need to succeed. Sign up here.