Mustafa Suleyman, cofounder of DeepMind, Google's AI group, believes that AI will be able to start and run its own firm within the next five years.
During a discussion on AI at the 2024 World Economic Forum, the now-CEO of Inflection AI was asked how long it will take AI to pass a Turing test-style exam. Passing would suggest that the technology has advanced to human-like capabilities known as AGI, or artificial general intelligence.
In response, Suleyman stated that the modern version of the Turing test would be to determine whether an AI could operate as an entrepreneur, mini-project manager, and creator capable of marketing, manufacturing, and selling a product for profit.
He seems to expect that AI will be able to demonstrate those business-savvy qualities before 2030—and inexpensively.
"I'm pretty sure that within the next five years, certainly before the end of the decade, we are going to have not just those capabilities, but those capabilities widely available for very cheap, potentially even in open source," Suleyman stated in Davos, Switzerland. "I think that completely changes the economy.”
The AI leader's views are just one of several forecasts Suleyman has made concerning AI's societal influence as technologies like OpenAI's ChatGPT gain popularity.
Suleyman told CNBC at Davos last week that AI will eventually be a "fundamentally labor-replacing" instrument.
In a separate interview with CNBC in September, he projected that within the next five years, everyone will have AI assistants that will enhance productivity and "intimately know your personal information.”
"It will be able to reason over your day, help you prioritise your time, help you invent, be much more creative," Suleyman stated.
Still, he stated on the 2024 Davos panel that the term "intelligence" in reference to AI remains a "pretty unclear, hazy concept." He calls the term a "distraction.”
Instead, he argues that researchers should concentrate on AI's real-world capabilities, such as whether an AI agent can communicate with humans, plan, schedule, and organise.
People should move away from the "engineering research-led exciting definition that we've used for 20 years to excite the field" and "actually now focus on what these things can do," Suleyman advised.