Google DeepMind CEO: AGI Is Coming ‘in a Few Years’

For all the wonders of ChatGPT and generative AI, these forms of artificial intelligence in their current form remain a far cry from the Holy Grail of the field: Artificial General Intelligence, or AGI.

AGI is achieved when intelligent machines mimic general human cognitive capabilities to perform any task, even if it is not trained on it. In contrast, today’s AI is adept at performing typically narrowly defined tasks.

For a long time, AGI is seen as many decades away from fruition, if not impossible to achieve.

That perception may be changing.

“We could be just a few years, maybe within a decade, away” to AGI, said Google DeepMind CEO Demis Hassabis, at The Wall Street Journal’s Future of Everything Festival. “The progress in the last few years has been pretty incredible. … I don’t see any reason why that progress is going to slow down. I think it may even accelerate.”

“I think we’ll have very capable, very general systems in the next few years,” he added.

His echoes the views of Geoffrey Hinton, whose seminal work on deep learning earned him a Turing award, computing’s equivalent to a Nobel Prize. This week, Hinton officially left Google so he could more freely warn about AI’s existential threat to humanity.

Hinton’s view of AI changed after he saw the power of GPT-4, OpenAI’s latest large language model which powers ChatGPT+. “Look at how it was five years ago and how it is now. Take the difference and propagate forwards. That’s scary,” he said.

Related:AI Takes Center Stage at 2023 RSA Conference: A Roundup of New Products

“The idea that this stuff could actually get smarter than people – a few people believed that,” Hinton told The New York Times. “But most people thought it was way off. And I thought it was way off. I thought it was 30 to 50 years or even longer away. Obviously, I no longer think that.”

This article originally appeared on AI Business.