During an interview at the Aspen Ideas Festival on Tuesday, Microsoft AI CEO Mustafa Suleiman made it clear that he respects OpenAI CEO Sam Altman.
CNBC's Andrew Ross Sorkin, using the metaphor of winning a bicycle race, asked what Microsoft's plans are for the AI giant once its future is less dependent on OpenAI, but Suleiman avoided answering.
“I don't believe in the metaphor of there being a finish line. That's another false framing,” he said. “We have to stop looking at everything as a blistering race.”
He then followed Microsoft's corporate lead with its deal with OpenAI, which is said to have invested $10 billion in a combination of cash and cloud credits. The deal gives Microsoft a large stake in OpenAI's commercial business, allowing it to incorporate the company's AI models into Microsoft products and sell the technology to Microsoft cloud customers. Some reports say Microsoft is also entitled to some payments from OpenAI.
“We're in fierce competition with them, for sure,” Suleiman said of OpenAI. “They're an independent company. We don't own them or control them. We don't even have board seats. So they're doing their own thing. But we have a deep partnership. I get on very well with Sam and have a lot of respect and trust and belief in what they've accomplished, and I expect that to continue for many years to come.”
This close-but-long relationship matters to Mr. Suleiman. Microsoft's investors and enterprise customers value it. But regulators are curious, and in April the EU acknowledged that the company's investments weren't true acquisitions. If that changes, regulators' involvement is likely to change, too.
Suleiman says he trusts Altman when it comes to AI safety
In a sense, Suleiman was the pre-OpenAI Sam Altman of the AI world: He spent most of his career competing with OpenAI and is known for his own ego.
Suleiman was the founder of AI pioneer DeepMind, which he sold to Google in 2014. In 2019, Bloomberg reported that he was placed on administrative leave following allegations of bullying employees, and he was subsequently transferred to other roles at Google before leaving the company in 2022 to join Greylock Partners as a venture partner. A few months later, he and Reid Hoffman, a Microsoft director at Greylock, founded Inflection AI, with a goal of, among other things, building their own LLM chatbot.
After Microsoft CEO Satya Nadella tried and failed to hire Sam Altman last fall, OpenAI fired him but reinstated him shortly after. Microsoft then hired Suleiman and most of Inflection in March, leaving the company in tatters and a big check. In his new role at Microsoft, Suleiman is auditing OpenAI's code, Semafor reported earlier this month. As one of OpenAI's former biggest foes, he now has the opportunity to take a deep dive into the inner workings of this arch-rival.
There's yet another wrinkle in all of this: OpenAI was founded on the premise that AI safety research would be used to stop evil AI from destroying humanity one day. In 2023, while still a competitor of OpenAI, Suleiman co-authored with researcher Michael Bhasker a book called “The Coming Wave: Technology, Power and the 21st Century's Greatest Dilemma,” which discusses the dangers of AI and how to prevent them.
Earlier this month, a group of former OpenAI employees signed a letter expressing concerns that OpenAI and other AI companies are not taking safety seriously enough.
When asked about it, Suleiman expressed his love and trust for Altman, but also said he wanted both regulation and a gentler change of pace.
“Maybe it's because I'm British and European, but I'm not afraid of regulation, which everyone rightly fears,” he said, describing the accusations from former employees as a “healthy conversation.” He added: “I think it's great that technologists, entrepreneurs and CEOs like myself and Sam are talking about regulation, someone I really love and think is amazing.” “He's not cynical, he's sincere. He believes it wholeheartedly.”
But, he also said, “friction is going to be your ally here. These technologies are going to be so powerful, so intimate, and so ever-present that now is the right time to get a grip on the situation.” Even if all these conversations slow down AI development by six to 18 months or more, “it's time well spent.”
There is a very close atmosphere among the players.
OpenAI CEO Sam Altman Image courtesy of TechCrunch
President Suleiman wants to cooperate with China and use AI in classrooms
Suleiman also made some interesting comments on other issues. On the AI race with China:
“With all due respect to our friends in Washington and the military-industrial complex, if the default framework is that this has no choice but to become a new Cold War, then that's exactly what it will be, because it becomes a self-fulfilling prophecy. They have no choice but to become hostile because they are afraid that we will become hostile, and this will only escalate,” he said. “We have to find ways to cooperate and be respectful of them, but at the same time acknowledge that we have different values.”
At the same time, he said, “China is building its own tech ecosystem and spreading it around the world. We should really be watching closely.”
When asked about his thoughts on kids using AI for school work, Suleiman, who doesn't have children, shrugged: “I think we need to be a little careful about fearing the downside of any tool. It's like when the calculator came out and there was this instinctive reaction of, oh no, anyone is going to be able to solve all the equations in a flash, and we're going to be stupider because we can't do mental arithmetic.”
He also predicts that as AI's linguistic capabilities improve, we could soon see AI acting as a sort of teacher's assistant, chatting live in classrooms: “What would it be like for a great teacher or educator to have an in-depth conversation with an AI in front of a live audience?”
The bottom line is that we may have unrealistic expectations if we want those building and profiting from AI to govern and protect humanity from its worst effects.