The run-of-show had all been carefully curated at Microsoft’s executive briefing center in Redmond, Wash., yesterday—where the tech company revealed the much-anticipated A.I.-powered version of its Bing search engine.
By harnessing OpenAI’s generative A.I. technology and pairing it with a plethora of its own data, Microsoft hopes it will finally have a chance to surpass Google in search. The new Bing, which the company started rolling out to users on a limited basis Tuesday, can recommend a five-day itinerary in Mexico City, compare flat-screen TVs, write source code, compose emails or social media posts, and translate messages into other languages.
“We're going to reimagine the search engine, the web browser, and new chat experiences,” Yusuf Mehdi, Microsoft’s corporate vice president and consumer chief marketing officer, said during the company’s presentation on Tuesday.
Microsoft executives uniformly stuck to the script in their comments, and carefully avoided discussing details like the financial cost of running the new Bing, or the specific version of GPT technology being used. Even Microsoft CEO Satya Nadella and Mehdi—both sporting navy pullover sweaters as they spoke to the media—seemed to have planned a matching wardrobe in advance.
Not everything went as planned. On Monday, OpenAI CEO Sam Altman tweeted a picture of himself and Microsoft’s Nadella, saying he was “excited for the event tomorrow”— and in so doing, alerting his 1.3 million followers of the launch event ahead of Microsoft’s carefully orchestrated schedule. Altman’s tweet prompted a response from Microsoft’s communications team, informing attendees that, “in light of recent announcements,” they could now talk freely about the upcoming briefing.
The new A.I. version of Bing, which Microsoft demoed to reporters after the keynote presentation, had a few mishaps: When asked about the death toll from the earthquake in Turkey and Syria, Bing cited data from Feb. 8, even though it was only Feb. 7, and Bing’s chatbot also stumbled through an attempt to identify authors on articles, saying it wasn’t clear from the website, even when it was.
All of this is exactly what you might expect to happen when a $2 trillion tech giant joins hands with an eight-year-old startup run by a Silicon Valley hotshot accustomed to giving his presentations in jeans. And it underscores how Microsoft is diverging from its competitors by tying itself so closely—and investing so much cash—in a startup. Microsoft may hold a lot of sway over OpenAI and its future revenues, and it may be its exclusive cloud provider, but it doesn’t have the autonomy to restrict OpenAI from working with other companies that directly compete against its own Microsoft Office or PowerPoint applications. And it sure can’t maintain absolute control of the narrative.
Altman—the only member of OpenAI’s team to present at Microsoft’s headquarters on Tuesday—spoke for just around two minutes, reading some of the same words from a teleprompter that appeared in a Microsoft blog post. But while he kept his comments to a minimum on stage, Altman’s past comments about A.I. are probably not the kinds of things Microsoft’s PR department would endorse: He has openly contemplated the dangers of revenge porn created by open-source image generators, and has said the worst-case scenario for A.I. is “lights-out for all of us.”
It’s a striking contrast to Microsoft’s own A.I. executives, who choose their words carefully when it comes to the potential dangers of A.I. During a panel at the event specifically about “responsible A.I.,” Microsoft’s chief responsible A.I. officer Natasha Crampton said that employees can flag concerns to a “sensitive use” team of researchers, engineers, and policy employees that deals with instances where the A.I. could have a consequential impact on someone’s life (like decisions on health care, education, or finance), could cause physical or psychological harm, or infringe on human rights in some way. But she wouldn’t specify how large that team was, and she wouldn't go into much more detail on how it operated.
OpenAI’s biggest benefit for Microsoft may come down to speed and urgency—something Nadella emphasized would be critical as the partners bring about the “new paradigm” for search.
“Rapid innovation is going to come,” Nadella said during the briefing. “In fact, that race starts today in terms of what you can expect. And we're going to move. We're going to move fast.”
Both Google and Microsoft appear to be dueling for first mover’s advantage—though it’s unclear who will ultimately have it. While OpenAI debuted its own ChatGPT model to the public at the end of last year, Microsoft only began to issue a limited release this week. Its models are initially quite impressive, but it remains to be seen how Google’s rival A.I. chat offering, Bard, will compare—and the exact point in time when both chatbots will become widely available search engine tools.
But the already-rapid pace at which both companies are trying to deploy their A.I. chatbots has some industry skeptics concerned—and raises questions as to whether there has been enough time to thoughtfully consider the gravity of potential safety issues at hand.
As for Nadella, the answer to that is clear.
“When you are talking about A.I., it’s about alignment with human preferences and societal norms,” Nadella said. “And you don't want to do that in a lab. You have to do that out there in the world.”
See you tomorrow,
Jessica Mathews
Twitter: @jessicakmathews
Email: jessica.mathews@fortune.com
Submit a deal for the Term Sheet newsletter here.
Jackson Fordyce curated the deals section of today’s newsletter.
