Anthropic CEO predicts human-level AI by 2026

Advances in artificial intelligence reaching “human levels” could occur as early as 2026, according to Anthropic CEO Dario Amodei.

In a conversation with podcaster Lex Fridman, Amodei discussed the rapid progression in artificial general intelligence (AGI) development and compared these advancements to various stages of education.

“We’re starting to get to PhD level, and last year we were at the undergraduate level, and the year before, the level of a high school student,” Amodei said.

He added that if the current pace continues, AI could exceed human cognitive abilities within the next few years.

“If you just eyeball the rate at which these capabilities are increasing, it does make you think that we’ll get there by 2026 or 2027,” he noted.

Amodei acknowledged that factors such as data limitations, scaling constraints, or geopolitical issues affecting microchip production could potentially slow this progress.

However, he stated that “if you believe the straight line extrapolation” of current growth rates, human-level AI could indeed arrive by 2026 or 2027, albeit with possible “mild delay.”

“Things that are powerful can do good things, and they can do bad things,” Amodei remarked, emphasising the need for responsible development.

“With great power comes great responsibility,” he added.

Anthropic, known for its large language model chatbot Claude, is among the companies competing with OpenAI's ChatGPT.

On the company’s approach, Amodei highlighted their “race to the top” mission, aiming to influence other industry players positively by setting examples.

When asked about the release date for the next iteration of Anthropic’s AI, Claude Opus 3.5, Amodei did not specify a timeline.

OpenAI CEO Sam Altman has echoed similar predictions, suggesting that AGI could be achieved within a few years with current technology.

AGI represents artificial intelligence that matches or surpasses human abilities across diverse cognitive tasks.