Dario Amodei, CEO of Anthropic, predicts that artificial intelligence could reach human-level intelligence by 2026. He made this statement during a podcast with Lex Fridman.

“We are approaching the level of a PhD. Last year, AI was at the undergraduate level, and the year before, it was like a high school student,” Amodei remarked, comparing AI’s progress to educational stages.

He believes that superintelligent AI could emerge as early as 2026 or 2027, though delays are possible.

Factors Slowing AI Development

Amodei highlighted several factors that could hinder AI progress:

Lack of high-quality data.
Challenges in scaling computational clusters.
Geopolitical conflicts impacting chip production.

Responsibility in the Age of Superintelligent AI

Amodei stressed the potential consequences of creating superintelligent AI, noting that such powerful entities could be used for both beneficial and harmful purposes.

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

Predictions from Other Experts

Previously, the CEO of OpenAI stated that humanity is on the verge of an era of unimaginable prosperity, predicting the advent of superintelligence within “a few thousand days.”

Current Limitations in AI

Some experts argue that AI has hit a plateau in its development due to the exhaustion of high-quality data for training, posing a challenge to further breakthroughs.