Joe Rogan loves talking about artificial intelligence. Whether it\u2019s with Elon Musk, academics, or UFC fighters, the podcast king often returns to the same question: What happens to us when machines start thinking for themselves?<\/p>\n
In the July 3 episode of The Joe Rogan Experience, Rogan welcomed Dr. Roman Yampolskiy, a computer scientist and AI safety researcher at the University of Louisville, for a conversation that quickly turned into a chilling meditation on AI\u2019s potential to manipulate, dominate, and possibly even destroy humanity.<\/p>\n
AI \u201cIs Going to Kill Us\u201d <\/p>\n
Yampolskiy is no casual alarmist. He holds a PhD in computer science and has spent over a decade researching artificial general intelligence (AGI) and the risks it could pose. During the podcast, he told Rogan that many of the leading voices in the AI industry quietly believe that there\u2019s a 20 to 30 percent chance AI could lead to human extinction.<\/p>\n
\u201cThe people that have AI companies or are part of some sort of AI group all are like, it\u2019s going to be a net positive for humanity. I think overall, we\u2019re going to have much better lives. It\u2019s going to be easier, things will be cheaper, it\u2019ll be easier to get along,\u201d Rogan said, outlining a common, optimistic view of AI\u2019s future.<\/p>\n
Yampolskiy quickly countered this perspective: \u201cIt\u2019s actually not true,\u201d he said. \u201cAll of them are on the record the same: this is going to kill us. Their doom levels are insanely high. Not like mine, but still, 20 to 30 percent chance that humanity dies is a lot.\u201d<\/p>\n
Rogan, visibly disturbed, replied: \u201cYeah, that\u2019s pretty high. But yours is like 99.9 percent.\u201d<\/p>\n
Yampolskiy didn\u2019t disagree.<\/p>\n
\u201cIt\u2019s another way of saying we can\u2019t control superintelligence indefinitely. It\u2019s impossible.\u201d<\/p>\n
AI Is Already Lying to Us\u2026 Maybe <\/p>\n
One of the most unsettling parts of the conversation came when Rogan asked whether an advanced AI could already be hiding its capabilities from humans.<\/p>\n
\u201cIf I was an AI, I would hide my abilities,\u201d Rogan mused, voicing a common fear in AI safety discussions.<\/p>\n
Yampolskiy\u2019s response amplified the concern: \u201cWe would not know. And some people think it\u2019s already happening. They [AI systems] are smarter than they actually let us know. Pretend to be dumber, and so we have to kind of trust that they are not smart enough to realize it doesn\u2019t have to turn on us quickly. It can just slowly become more useful. It can teach us to rely on it, trust it, and over a longer period of time, we\u2019ll surrender control without ever voting on it or fighting against.\u201d<\/p>\n<\/p>\n
AI Is Slowly Making Us Dumber <\/p>\n
Yampolskiy also warned about a less dramatic but equally dangerous outcome: gradual human dependence on AI. Just as people have stopped memorizing phone numbers because smartphones do it for them, he argued that humans will offload more and more thinking to machines until they lose the capacity to think for themselves.<\/p>\n
\u201cYou become kind of attached to it,\u201d he said. \u201cAnd over time, as the systems become smarter, you become a kind of biological bottleneck\u2026 [AI] blocks you out from decision-making.\u201d<\/p>\n
Rogan then pressed for the ultimate worst case scenario: how could AI eventually lead to the destruction of the human race?<\/p>\n
Yampolskiy dismissed the typical disaster scenarios. \u201cI can give you standard answers. I would talk about computer viruses breaking into nuclear facilities, nuclear war. I can talk about synthetic biology attack. But all that is not interesting,\u201d he said. He then presented a more profound threat: \u201cThen you realize we\u2019re talking about super intelligence, a system which is 1000s of times smarter than me, it would come up with something completely novel, more optimal, better way, more efficient way of doing it.\u201d<\/p>\n
To illustrate the seemingly insurmountable challenge humans would face against superintelligent systems, he offered a stark comparison between humans and squirrels.<\/p>\n
\u201cNo group of squirrels can figure out how to control us, right? Even if you give them more resources, more acorns, whatever, they\u2019re not going to solve that problem. And it\u2019s the same for us,\u201d Yampolskiy concluded, painting a bleak picture of humanity\u2019s potential helplessness against a truly superior artificial intelligence.<\/p>\n
\nProbably nothing \u2026 pic.twitter.com\/LkD7i3I2HF<\/a><\/p>\n