AI could soon run wild, warns Anthropic's chief scientist
Jared Kaplan, Anthropic's chief scientist, says we're heading for a big decision about AI between 2027 and 2030: should we let AI train itself?
If we do, he warns it might trigger an "intelligence explosion," creating super-smart systems that could be tough to control.
Kaplan states, "That's the thing that we view as maybe the biggest decision or scariest thing to do... once no one's involved in the process, you don't really know," echoing concerns from other experts who fear things could get out of hand.
What this means for jobs (and why humans still matter)
Kaplan predicts AI will soon take over most white-collar jobs—a view shared by leaders at OpenAI and Anthropic.
He stresses that to avoid losing control or facing unexpected problems, it's crucial for humans to stay in charge of how these powerful AIs are developed and used.