Here's how far we are from AGI, according to the people developing it
- AGI, or artificial general intelligence, is a still-theoretical AI that can reason like humans.
- Top researchers agree the leap to AGI is close but differ on just how close.
- Some say we'll see AGI in as little as two years. Others say we're still decades away.
One of the oft-stated goals of the current AI arms race is to reach artificial general intelligence, or AGI.
There is no agreed-upon definition. Generally, it's a hypothetical form of machine intelligence that can solve any human task through methods that aren't constrained to its training.
OpenAI CEO Sam Altman, a strong proponent of AGI, has said it will "elevate humanity." He envisions it as a tool to enhance productivity, create shared intelligence, increase abundance, and discover new knowledge.
Others have likened it to intelligence smarter than humans or even a "God-like AI" that could make humans obsolete.
However, the question of when we'll reach it is still a debate among many of the top names in the field. Here's a closer look at how far we are from AGI, according to the people closest to it.
The CEO of OpenAI believes that we're already making major progress toward AGI. On the Y Combinator podcast, he said one of the things he's most excited about in 2025 is the arrival of AGI.
OpenAI's former head of AGI readiness believes we'll see some form of AGI manifest in the next few years. Brundage, who left OpenAI in August, said that in the next few years, the AI industry will develop "systems that can basically do anything a person can do remotely on a computer." That includes operating the mouse and keyboard or even looking like a "human in a video chat."
Dario Amodei, CEO of OpenAI's competitor Anthropic, believes we will see some form of AGI by 2026. In an essay he posted to the company's website in October, he described AGI, which he prefers to call "powerful AI," as smarter than a Nobel Prize winner across many fields, multimodal, independent, fast, replicable, cooperative, and free of a physical embodiment. In short, he believes it'll be akin to "a country of geniuses in a data center."
AI godfather Geoffrey Hinton believes that we might see AI that's smarter than humans in as little as five years.
In a post on X last year, he wrote, "I now predict 5 to 20 years but without much confidence. We live in very uncertain times."
Demis Hassabis, the CEO of Google DeepMind and a recently minted Nobel laureate, believes that an AI that can reason as well as humans is still a decade away, at least.
Leading AI researcher Andrew Ng is a little more conservative in his estimates about AGI. He's described it as a form of intelligence that can do "any intellectual tasks that a human can," whether that's driving a car, flying a plane, or writing a Ph.D. thesis.
And he's not convinced we'll get there soon. "I hope we get there in our lifetime, but I'm not sure," he said, adding that people should be skeptical of companies that claim AGI is imminent.
Richard Socher, a former Salesforce executive who is now the CEO of AI-powered search engine You.com, says there are two ways to define AGI. "There's a simple economic one, which is 80% of the jobs will be automated with AI, and then we can call it AGI," he previously told Business Insider. He predicted we'll get there in three to five years.
When you expand the definition of AGI to a form of intelligence that can "learn like humans" and "visually have the same motor intelligence, and visual intelligence, language intelligence, and logical intelligence as some of the most logical people," then the timeline could range from as little as 10 years to as much as 200 years, he said.
Meta's chief AI scientist, Yann LeCun, doesn't believe we'll see AGI anytime soon.
At the World Economic Forum's annual meeting in Davos in January, LeCun said AGI "is not around the corner" and it "will take years, if not decades" before it comes to fruition.
And LeCun says we shouldn't expect it to arrive as a single event.
On an episode of Lex Fridman's podcast in March, he said, "The idea somehow which, you know, is popularized by science fiction and Hollywood that, you know, somehow somebody is going to discover the secret, the secret to AGI, or human-level AI, or AMI, whatever you want to call it. And then, you know, turn on a machine, and then we have AGI. That's just not going to happen. It's not going to be an event. It's going to be gradual progress."