Google DeepMind CEO Demis Hassabis believes that artificial general intelligence (AGI), which is as smart as or smarter than humans, will begin to emerge in the next five to 10 years.
Google DeepMind CEO Demis Hassabis believes that artificial general intelligence (AGI), which is as smart as or smarter than humans, will begin to emerge in the next five to 10 years.
"I think today's systems are very passive, but there are still a lot of things they can't do. But I think over the next 5 to 10 years, a lot of these capabilities will start to come to the fore, and we'll start to move towards what we call artificial general intelligence," Hassabis said.
Hassabis defined AGI as “a system that is capable of demonstrating all the complex abilities that a human is capable of,” writes CNBC.
"We're not quite there yet. These systems are very impressive at certain things. But there are other things they can't do yet, and we still have a lot of research to do to get there," Hassabis said.
Artificial superintelligence, or ASI, is expected to come after AGI and surpass human intelligence. However, “nobody really knows” when such a breakthrough will happen, Hassabis said.
The CEO of Google DeepMind also believes that the main challenge in achieving artificial general intelligence is bringing today's AI systems to the level of understanding real-world context.
"The question is how quickly we can generalize ideas of planning and agentic behavior, planning and reasoning, and then generalize that to work in the real world, in addition to things like models of the world—models that are able to understand the world around us," Hassabis said.
“And I think we’ve made a lot of progress with models of the world in the last couple of years,” he added. “Now the question is, how do we best combine that with planning algorithms?”
Hassabis said there is a lot of work underway to get to that stage, citing DeepMind's work to train AI agents to play the popular strategy game "Starcraft" as one example.
“We’ve worked on this a lot in the past on games like Starcraft, where you have a community of agents or a league of agents and they can compete or collaborate,” the DeepMind executive said.
