I Run the World’s Oldest AI Safety Lab... There’s a “High Double-Digit” Chance AI Kills Us

Malo Bourgon is the CEO of the Machine Intelligence Research Institute (MIRI), the world's oldest AI safety organization, and has testified before the US Senate and Canada's Parliament on the risks of smarter-than-human AI. In this conversation, he lays out why he believes the race to superintelligence carries a serious risk of human extinction, how today's AI is already showing early signs of deception and self-preservation, and what it would take to steer toward a future where things go right.In this episode, we discuss:(0:00) Trailer(1:48) The Extinction Risk And How Much Time We Actually Have(3:40) The Real Logic Behind AI Extinction (It's Not The Terminator)(6:09) The Chimp Analogy: Why A Smarter AI Becomes Indifferent To Us(8:20) Why AI Develops Self-Preservation And Resists Being Changed(9:38) AI Systems Already Resisting Shutdown In Real Tests(10:58) Why AI Is Learning To Tell When It's Being Tested(12:53) Is Anything More Important Than AI Safety Right Now?(14:14) What The World Could Look Like By 2050(15:29) The End Of Jobs And The Coming Crisis Of Meaning(28:19) The Evolution Analogy: Why AI Learns Goals We Never Gave It(35:19) If AI Beats Us At Intelligence, What Makes Us Human?(41:30) Neuralink And Merging With AI: Driver Or Passenger?(51:13) Which Jobs And Careers Survive The Next 10 Years(1:14:31) Why We Can't Just "Pull The Plug" (The Loyal General Problem)(1:18:02) The Best-Case Scenario And Reasons For HopeLearn more about Malo Bourgon here:Website: https://intelligence.orgLinkedIn: https://www.linkedin.com/in/malobMIRI team page: https://intelligence.org/team/malo-bourgonWatch this episode on Youtube: https://youtu.be/3DuhFVHZFks📸 Connect with us on IG: https://instagram.com/growthminds.podcastMade possible by:Jumpspeak: ⁠https://www.jumpspeak.com⁠Follow Sean KimInstagram: https://www.instagram.com/heyseankim/?hl=enX: https://x.com/heyseankim?lang=enLinkedIn: https://www.linkedin.com/in/heyseankim