The race to create artificial general intelligence, or AGI, has become a major focus for the tech giants. But Daniela Amodei, president and cofounder of Anthropic, says the term itself may no longer be the best way to think about AI progress. “AGI is such a funny term,” Amodei told CNBC. “Many years ago, it was kind of a useful concept to say, ‘When will artificial intelligence be as capable as a human?'” She explained that this idea is starting to lose its meaning. In some areas, AI systems have already matched or even exceeded human abilities. For example, Anthropic’s Claude model can write computer code at a level similar to many professional engineers, including some working at the company.
Yet, she noted that AI still struggles with tasks that humans find simple. “Claude still can’t do a lot of things that humans can do,” she said, highlighting the gap between AI performance in specific tasks and overall human intelligence.
Also read: CES 2026: Google and XREAL deepen Android XR partnership, check details
Because of this, Amodei believes the concept of AGI may be outdated or at least no longer useful as a measure of progress. “I think maybe the construct itself is now wrong- or maybe not wrong, but just outdated,” she said.
As companies like Anthropic invest billions of dollars in larger models and data centers, some critics worry that AI might not achieve true general intelligence without major breakthroughs. Amodei said, “We don’t know,” she said of what breakthroughs may still be needed. “Nothing slows down until it does.”
Also read: Man claims AI girlfriend dumped him after he spoke against feminism: Here’s what happened
Instead of focusing on whether AI reaches AGI, Amodei emphasised looking at how these systems are used in the real world.
For Amodei, the real question is not if AI becomes human-level in every sense, but what it can do, where it falls short, and how society chooses to use it.