Join us in returning to NYC on June 5th to collaborate with executive leaders in exploring comprehensive methods for auditing AI models regarding bias, performance, and ethical compliance across diverse organizations. Find out how you can attend here.
It’s a big week for Mustafa Suleyman. Microsoft’s recently installed AI division head will be front and center at the company’s Build developer conference. With OpenAI and Google having both announced new models and AI programs, all eyes now turn to Seattle, Washington, to see how the Windows maker will respond.
Ahead of his first Build, Suleyman sat down with Microsoft’s Chief Communications Officer Frank Shaw to discuss everything about AI.
Copilot everywhere
One of Build’s major themes is the evolution of Copilot, Microsoft’s Cortana replacement that has become infused into every aspect of the company’s product family. It’s integrated with Microsoft Teams, GitHub, Power Pages, 365, and Windows. But its assimilation isn’t what surprises Suleyman; it’s the appropriation by other companies—Copilot has become the AI version of “Xerox.”
“The cool thing about the brand Copilot is that it is really starting to define the category,” he tells Shaw. “Copilot isn’t just a brand that we own. Copilot is now becoming a phrase that people use to refer to AI assistants, personal AIs, productivity AIs, work AIs, and many companies…and startups are starting to use it. So that’s actually been quite surprising to me…I got a good feeling about all the different companies using that phrase. And I think that’s a real testament.”
VB Event
The AI Impact Tour: The AI Audit
Join us as we return to NYC on June 5th to engage with top executive leaders, delving into strategies for auditing AI models to ensure fairness, optimal performance, and ethical compliance across diverse organizations. Secure your attendance for this exclusive invite-only event.
Suleyman describes Copilot as a “confident name” because it acknowledges that AI will live with you in the background, helping and supporting you throughout your life. “That’s a very bold statement in itself. This isn’t just a new technology platform shift. This is going to be a profound shift in what it means to be human and how we interact with our tools.”
He claims that until now, the tools created by technology have been “relatively static,” meaning they can be turned on and off but require manual oversight and control. However, in an AI world, more “dynamic, emergent, and interactive experiences” will be a part of our everyday lives.
AI needs EQ
AI development is often shaped by two sides: proponents of its progression and those who urge caution and warn about the technology’s dangers to society. Concerns by so-called doomsayers may take on renewed meaning now following OpenAI’s dismantling of the team tasked with ensuring superintelligent AI models don’t run amuck. But Suleyman offers a different perspective, explaining the terms “safety” and “ethics” have become “a bit tired and overused over the years.” Instead, AI needs to be infused with emotional intelligence.
“The thing that I’ve been focused on in the last two or three years is what does it mean to create experiences that are emotionally intelligent, not just functional productivity tools that can do reasoning, follow instructions and provide you with logical or factual answers on their own,” he remarks. “Most people really care about the connection they have with a brand, the feeling, the tone, energy they get by the kind of color palette, the words, and the experience design.”
He realized this years ago with his former startup, Inflection AI before Microsoft hired him and much of his team away. Suleyman doesn’t advocate for an elimination of guardrails. Instead, he suggests creating an emotional experience between us, and the AI tools we use could be a better way forward. It could be something as simple as having AI ask how we’re feeling, what we’re thinking about, how we’re processing our day, or “just being genuinely even-handed about a topic…encouraging you to empathize with someone or some idea you might be unfamiliar with or might disagree with.”
One way to incorporate humanity into AI is by allowing models to not be “too verbose” and allowing them to embrace sensitive issues. “If you just shut somebody down when they have a contrarian or sort of somewhat irregular ‘view’ with respect to minorities, abortion, immigration or take your sensitive cultural issue of the day, dialing people down is one of the worst things we could possibly do because that just compounds the judgment and polarization that is driving this mainstream media versus outside media,” Suleyman argues. “And I think that’s a really important thing for us as an organization to lean into: subtlety, nuance, respect, genuine empathy for those we fundamentally might disagree with.”
It’s an interesting perspective, but having it embraced by the public is easier said than done. One need only look at what happened with Microsoft’s Tay chatbot, reactions to Google’s diversity issue with its image generation tool, and more. But Suleyman asserts that EQ, to some degree, is already baked into AI and promotes it as a great time to create because we’re seeing tools that are responsive to this kind of design.
Setting the stage for 1,000 Einsteins
Suleyman has a storied history with AI. Prior to Microsoft and Inflection, he co-founded DeepMind before its acquisition by Google in 2014. So, what fuels his passion?
“I’ve always been attracted to ways to change the world,” he explains to Shaw while admitting his answer sounds “super cheesy.” He goes on to say, “I have always been attracted to the idea that we could reduce the cost of accessing super high-quality personalized information to basically zero marginal cost.” In addition, Suleyman believes we can also “reduce the cost of accessing kindness and support to almost zero.”
Microsoft’s AI chief predicts that if AI provides this kind of support and knowledge to everybody, society will produce the “next 1,000 Einsteins that help us unlock all of our biggest social challenges that we’re facing in climate, food, healthcare, education, and so on for the next century.”
It’s not difficult to see hints of this vision already in existence. Look through many AI announcements, and you’ll get a sense the goal is to remove mundane and tedious tasks from the human workload so that people can embrace their creative side. Doing so could free up resources to tackle the aforementioned social challenges.
The future computer
Overall, Suleyman provides a compelling vision of the future of AI development and the role Microsoft’s technology may play in it. To him, “we’ve been scratching the surface of what it means to have an ever-present super smart, very reliable, stable, factually-grounded Copilot” in our lives.
The future computer will understand everything we say, all the time, including what we’re thinking and feeling. “It’s not going to look like the current shape of computing,” he opines. “We will still have laptops, tablets and desktops—they’re still going to be a big part of what we do, but there is going to be a new set of surfaces which feel quite different.”
One need only look at the nascent AI PC market and the AI-powered wearable space. They remind us that we’re in for a big change in how we use technology.
