Microsoft is building an “intimate experience” in AI assistants as Mustafa Suleyman, Microsoft AI CEO and co-founder of DeepMind, focuses Microsoft AI’s team on the business of engineering personality.

“We’re crafting a lasting, meaningful, trusted relationship,” Suleyman said during a fireside chat at the recent 2024 IA Summit. “That is the new platform as far as I see it because it’s not just about voice and text or language — it’s really going to be about vision.”

AI companions will see everything that we see in our browser and on our desktop in real-time, understanding both text and images with the ability to talk about it easily.

Perhaps Recall will become a first step — and the foundation — to advance this technology.

Microsoft Recall is an AI search feature for Windows that takes screenshots of activity on Windows-run computer screens so people can later search their machine for specific information.

During the fireside chat, Suleyman called AI assistants an “ever-present, fully aware, perpetual memory companion” that creates an “intimate experience” like sitting on the couch at night with your partner and searching for something to do during the holidays.

Suleyman has begun to rethink the way Bing searches and AI work. In a post on X, he noted a focus on simplicity and commended the Bing team — which “unshipped 27 features” — and called the team “brave” for taking the initiative to “undo complexity” in the platform.

At the summit, Suleyman shared his vision for how personal AI assistants will transform the way people live, work and interact with technology.

Not having the ability to predict what the world would look like in the next three to five years is normal, Suleyman said.

“In the early days of the internet, it was super hard to predict that mobile would take the form that it did, and business models would adapt to this new modality,” Suleyman said.

He added: “The same is true now, even though we have a pretty clear line of sight for the capabilities that will emerge in the next couple of years, it’s difficult to say what the combination of capabilities means for existing ecosystems, for incentive structures and for the kinds of businesses that we’re going to build.”

He said that for the first time in human history, machines have learned to speak the same type of language as humans. It has fundamentally changed programming interfaces.

Humans are not effective at making predictions and yet they do it all the time, Suleyman said.

Still, he acknowledges, it is possible to predict capabilities the world will likely see in the next two or three years. The difficult task becomes translating those capabilities into new products, business models and ecosystems.

“If I were to bet on the capabilities, I would say, well, we are doing pretty well at IQ,” he said. “The models are increasingly factual [and] they’re reducing hallucinations really significantly. They can retrieve well and conditionally generate over arbitrary documents, whether the web or your private corpus.”

Gaining IQ is only one part of this. Empathy is another. Years ago, experts thought AI would never have “emotional intelligence.” He said that always will be the “special magic.”

“And yet I think they’re really starting to show very elegant, very graceful, very fluid interactions obviously in some of the answers they give in text, but also with the voice,” he said.

Memory is what will connect everything. Experts are working on that — taking different approaches — but I’m pretty sure in the next 18 months we’re going to have AIs with very, very good memory, if not infinite and perfect, but very good ability to retrieve over arbitrary documents,” he said.