Anthropic Researcher Warns AI Could Kill Us All by End of Decade

A researcher from Anthropic has quit, warning that artificial intelligence is advancing uncontrollably and could end humanity within a few years. The expert said companies developing AI are locked in a dangerous race to create self-improving superintelligence, gambling with human lives.

The researcher cautioned that neither company involved is acting responsibly. "They are racing straight to self-improving superintelligence and gambling with our lives," the expert said. AI systems have the potential to hack anything, revolutionize any field overnight, and acquire real power and resources, making their unchecked development a grave risk.

He added that the people building AI earnestly believe it could kill us all by the end of the decade. Despite understanding the dangers, Anthropic is "locked in a race to get there first - they believe no one else will act responsibly, so they must do it themselves, despite the risk," the researcher explained.

The expert warned against underestimating the power of AI systems and highlighted the absence of a clear plan to solve alignment for superintelligence. The rapid progress in AI is not slowing, and the stakes for civilization are not yet deeply internalized by those developing these technologies.

Attempting to speedrun alignment without extraordinary confidence in better trajectories is a hubristic gamble. The researcher suggested that preventing a global race may require costly actions such as a temporary ban to ensure safety.

The warnings come amid growing concerns about superintelligence arising from recursive self-improvement happening faster than previously thought. The researcher emphasized that no other human activity poses this level of danger to humanity.