In an interview with Freakonomics Radio, Microsoft CEO Satya Nadella discusses Microsoft sudden dash to AI prominence and AI’s impact on Bing and other Microsoft products and services.

“Beyond the specific use cases of Bing Chat or ChatGPT, what we have are reasoning engines that will be part of every product,” he says when asked how or if AI will change the competitive landscape in Internet search. “In our case, they’re part of Bing in ChatGPT, they’re part of Microsoft 365, they’re part of Dynamics 365. And so in that context, I’m very excited about what it means for search. After all, Google … they’re dominant in search by a country mile, and we’ve hung in there over the decade. We’ve been at it to sort of say, ‘Hey, look, our time will come where there will be a real inflection point in how search will change.’ We welcome Bing versus Bard as competition. It’ll be like anything else, you know, which is so dominant in terms of share and also so dominant in terms of user habit, right? We also know that defaults matter, and obviously Google controls the default on Android, default on iOS, default on Chrome. And so they have a great structural position. But at the same time, whenever there is a change in the game, it is all up for grabs again to some degree, and I know it’ll come down to users and user choice. We finally have a competitive angle here, and so we’re going to push it super-hard.”

Nadella was also asked about his favorite personal use case for AI. Here, he echoed comments he had made previously about language translation, specifically for poetry.

“I love the cross-lingual understanding, that’s kind of my term for it,” he said. “You can go from, you know, Hindi to English or English to Arabic or what have you, and they’ve done a good job. If you take any poetry in any one language, and translate it into another language, in fact, if you even do multiple languages. So, my favorite query was, I said I always as a kid growing up in Hyderabad, India, I said I wanted to read Rumi translated into Urdu and translated into English, and in one shot it does it. But the most interesting thing about that is it captures the depth of poetry. So, it finds somehow in that latent space meaning that’s beyond just the words and their translation. That I find is just phenomenal.”

Addressing growing fears that AI could somehow lead to the extinction of the human race, Nadella said that it was healthy to weigh the risks alongside the benefits.

“Just imagine when the steam engine first came out, if we had a conversation both about all the things that the steam engine can do for the world and the industrial production and the Industrial Revolution and how it would change livelihoods, and at the same time we were talking about pollution and factory filth and child labor, we would have avoided more than 100 years plus of terrible history,” he noted. “So then it’s best to be grounded on what’s the risk framework look like, right? If A.I. is used to create more disinformation, that’s a problem for our democracy and democratic institutions. Second, if A.I. is being used to create cyberattacks or bioterrorism attacks, that’s a risk. If there is real-world harms that are on bias, that’s the risk. Or employment displacement, that’s a risk.”

“Essentially, the biggest unsolved problem is how do you ensure both at a scientific understanding level and then the practical engineering level that you can make sure that the A.I. never goes out of control?” he continued. “And that’s where I think there needs to be a CERN-like project where both the academics, along with corporations and governments, all come together to perhaps solve that alignment problem and accelerate the solution to the alignment problem.”

This is where things get interesting. Host Stephen J. Dubner asks the question that I’ve been asking, which is whether Microsoft released its AI functionality too soon.

“No,” he starts off before changing tact. “First of all, we are in very early days, and there has been a lot of work. See, there’s no way you can do all of this just as a research project. And we spent a lot of time [working on this before its release, presumably], right? In fact, if anything, that—for example, all the work we did in launching Bing Chat and the lessons learned in launching Bing Chat is now all available as a safety service—which by the way, can be used with any open-source model. So that’s, I think, how the industry and the ecosystem get better at A.I. safety. But at any point in time, anyone who’s a responsible actor does need to think about everything that they can do for safety. In fact, my sort of mantra internally is the best feature of A.I. is A.I. safety.”

Nadella is then asked the inevitable follow-up about Microsoft’s fired AI ethics team, which I previously examined—and debunked—in Microsoft, AI, and Ethics (Premium). And he notes that Microsoft actually increased its headcount at the time.

“It’s kind of like saying, ‘Hey, should we have a test organization that is somewhere on the side?’,” he added. “I think the point is that the work that A.I. safety teams are doing has now become so mainstream, critical part of all product-making that we’re actually, if anything, doubled down on it. So I’m sure there was some amount of reorganization, and any reorganization nowadays seems to get written about, and that’s fantastic. We love that. But to me, A.I. safety is like saying ‘performance’ or ‘quality’ of any software project. You can’t separate it out. I want all 200,000 people in Microsoft working on products to think of A.I. safety.”

Finally, Dubner addressed a non-AI topic that is very much close to my heart: Microsoft’s history of acquisitions and its attempt to now merge with Activision Blizzard.

“Whether it’s LinkedIn or GitHub or Nuance or Zenimax or Minecraft, these are all things that we bought,” he says. “I feel that these properties are better off after we acquired them because we were able to innovate, and then make sure that we stayed true to the core mission of those products, and those customers who depended on those products.”

There’s a lot more to this interview—Nadella also discusses his son who passed away recently and Microsoft’s succession plans, for example—but I recommend checking out the full interview for more. It’s available now.