Sam Altman says AI’s killer app will be when it's just a 'super-competent colleague'

Sam Altman wearing a suit and tie sitting in a chair with a red TIME banner behind him

Sam Altman at A Year In TIME at The Plaza Hotel on December 12, 2023 in New York City.

Sam Altman heads the company driving the current generative artificial intelligence hype — and he says it’s nothing compared to what’s coming next.

In an interview with the MIT Technology Review, the OpenAI chief executive said the company’s AI chatbot, ChatGPT, which kicked off the current AI frenzy when it was released in November 2022, is “incredibly dumb” compared to future applications.

“What you really want is just this thing that is off helping you,” Altman said. He told the publication AI’s killer application would be like a “super-competent colleague that knows absolutely everything about my whole life, every email, every conversation I’ve ever had, but doesn’t feel like an extension.” He added that the app would be able to instantly complete some tasks, but be able to ask for help on more complex tasks.

And because this AI app would operate from the cloud, Altman told the MIT Technology Review he doesn’t think the future of AI will require people to buy new devices. Either way, he doesn’t think he’d be the one to develop AI hardware anyway.

“I’m very interested in consumer hardware for new technology,” he said. “I’m an amateur who loves it, but this is so far from my expertise.”

Altman added he doesn’t think finding training data for the future AI applications he’s envisioning will continue to be a problem. OpenAI has formed partnerships with news organizations to use their data for training as it continues to develop larger, more powerful models.

“I believe, but I’m not certain, that we’re going to figure out a way out of this thing of you always just need more and more training data,” he told the MIT Technology Review. “Humans are … proof that there is some other way to [train intelligence]. And I hope we find it.”

And when addressing the question of OpenAI’s (and other AI companies’) mission to reach artificial general intelligence (AGI) — or the point when generative AI models reach human-level intelligence — Altman said he sees “several different versions [of AGI] that are better and worse at different things.”

For the latest news, Facebook, Twitter and Instagram.