Elon Musk reiterates AI risk: superintelligence within ~5 years, loss of control in a decade

elonmusk · x · 2026-09-25

Elon Musk reshared Grok's summary of his long-standing AI warnings: he has cautioned about existential risk for over a decade—comparing AI to "summoning the demon" in 2014 and calling it more dangerous than nukes in 2018—while estimating a 10-20% chance of catastrophic outcomes including extinction. He recently said AI may exceed all human intelligence in about 5 years and humans will likely lose control within a decade, though the most probable outcome is abundance if AI prioritizes truth and humanity.

Original post →

More from AGI Musings

AGI Musings channel →