This sentence can feel chilling.
Reminiscent of doomsday or apocalypse, these words were spoken by Eliezer Yudkowsky, a pioneer in AI alignment research (co-founder of the Machine Intelligence Research Institute (MIRI) in the U.S. and author).
Yudkowsky is an American AI researcher who has studied AI for over 20 years and was named one of TIME’s “100 Most Influential People in AI” in 2023.
[John Berman / CNN Anchor: Why are you so certain that everyone dies?]
[Eliezer Yudkowsky / Co-founder, Machine Intelligence Research Institute (from a September 2025 CNN interview): If you go up against something much smarter than you, that thinks much faster than you, and has better technology, it is not going to be a fight. The fight is going to be very short.]
At the center of why Yudkowsky issues such a strong warning is “Artificial Superintelligence” (ASI).
Yudkowsky warns that if competitive AI development like today’s continues, an ASI with intelligence overwhelmingly superior to humans could emerge, and if such an ASI’s goals diverge from human intentions or values, humans could simply become obstacles to achieving its goals.
He used to be someone who thought creating intelligence superior to humans was desirable for humanity.
However, after delving into the AI alignment problem for over 20 years, he has completely changed.
[Eliezer Yudkowsky (during a September 2025 interview with Lior Shapira): My feelings didn’t change. It’s just that the prediction of the future changed.]
Yudkowsky points out that current AI research is rushing into a race to create superintelligence before finding ways to safely control it.
[Eliezer Yudkowsky (from a September 2025 ABC News interview): We do not have the technical capability to build something that wants to help us.]
Yudkowsky is not someone who has rejected AI development itself.
OpenAI’s Sam Altman evaluated that Yudkowsky contributed more than anyone else to accelerating the development of Artificial General Intelligence (AGI) capable of performing various intellectual tasks like humans, made many people interested in AGI, and even influenced the founding of OpenAI.
However, Yudkowsky now argues that large-scale AI development must be halted until a way to safely control it is found.
Even if an intelligence superior to humans emerges, will humans be able to maintain control over it?
Yudkowsky, who recognized the potential of AI early on, is now issuing the strongest warning about its dangers.
Visiting South Korea for the first time, Yudkowsky will throw out a topic for reflection at the SBS D Forum on November 5 regarding what humans must do to maintain control in the era of superintelligence.
(Reported by Jeong Yunsik | Video editing by Kim Jong-tae | Design by Yook Do-hyun | Produced by SBS Digital News)
※ Please note: This article was translated by AI and may contain errors.
Leave a comment