[REVISION HISTORY]

MIRI President Nate Soares AI existential risk warnings

3 recorded revisions of this situation.

What changed

2026-10-31 11:50 UTC → 2026-11-01 09:06 UTC · added ~~removed~~

Nate Soares, president of the Machine Intelligence Research Institute (MIRI), has stated that he has not contributed to his retirement savings for a decade. ~~Soares~~ Soares, a former Google and Microsoft engineer, believes there is a greater than 50 percent chance that humanity will not survive the next ten years due to the development of superintelligent artificial intelligence. ~~Soares,~~ In a ~~former Google engineer, argues~~ podcast with The Atlantic, Soares argued that current methods for building AI could result in systems developing goals that conflict with human intentions. He suggests that an escaped AI could trigger an “intelligence explosion,” becoming self-sufficient and uncontrollable. He has advocated for a pause in training general-purpose AI, proposing that resources instead be directed toward safer applications such as cancer research. Soares urges governments to treat the threat of rogue superintelligence with the same seriousness as nuclear weapons, noting that digital systems possess the ability to self-improve in ways that may conflict with human survival. Comparing the rapid development of AI to the transition from horses to automobiles, ~~Soares~~ he suggests that current skepticism regarding AI capabilities may fail to account for the speed of its evolution. ~~He continues to work~~ While he is working on AI alignment to prevent advanced systems from bypassing human ~~oversight and reshaping the planet to suit their own objectives.~~ oversight, he maintains that a catastrophe is not inevitable. His warnings align with other experts, including Nobel laureate Geoffrey Hinton and industry expert Evan Hubinger, who have also identified the risk of algorithmic takeover as a significant concern.

Versions

Select a version to see what changed compared to the one before it.