Joe Rogan’s Newest Episode Will Make You Query The whole lot About AI

Joe Rogan loves speaking about synthetic intelligence. Whether or not it’s with Elon Musk, teachers, or UFC fighters, the podcast king typically returns to the identical query: What occurs to us when machines begin considering for themselves?

Within the July 3 episode of The Joe Rogan Expertise, Rogan welcomed Dr. Roman Yampolskiy, a pc scientist and AI security researcher on the College of Louisville, for a dialog that rapidly was a chilling meditation on AI’s potential to control, dominate, and probably even destroy humanity.

AI “Is Going to Kill Us”

Yampolskiy is not any informal alarmist. He holds a PhD in laptop science and has spent over a decade researching synthetic normal intelligence (AGI) and the dangers it might pose. Throughout the podcast, he advised Rogan that most of the main voices within the AI trade quietly consider that there’s a 20 to 30 p.c likelihood AI might result in human extinction.

“The folks that have AI corporations or are a part of some kind of AI group all are like, it’s going to be a internet optimistic for humanity. I believe general, we’re going to have a lot better lives. It’s going to be simpler, issues will probably be cheaper, it’ll be simpler to get alongside,” Rogan mentioned, outlining a standard, optimistic view of AI’s future.

Yampolskiy rapidly countered this attitude: “It’s really not true,” he mentioned. “All of them are on the file the identical: that is going to kill us. Their doom ranges are insanely excessive. Not like mine, however nonetheless, 20 to 30 p.c likelihood that humanity dies is so much.”

Rogan, visibly disturbed, replied: “Yeah, that’s fairly excessive. However yours is like 99.9 p.c.”

Yampolskiy didn’t disagree.

“It’s one other method of claiming we are able to’t management superintelligence indefinitely. It’s inconceivable.”

AI Is Already Mendacity to Us… Possibly

One of the vital unsettling components of the dialog got here when Rogan requested whether or not a sophisticated AI might already be hiding its capabilities from people.

“If I used to be an AI, I might cover my skills,” Rogan mused, voicing a standard concern in AI security discussions.

Yampolskiy’s response amplified the priority: “We’d not know. And a few individuals suppose it’s already occurring. They [AI systems] are smarter than they really tell us. Fake to be dumber, and so we’ve to form of belief that they don’t seem to be sensible sufficient to understand it doesn’t must activate us rapidly. It could possibly simply slowly grow to be extra helpful. It could possibly educate us to depend on it, belief it, and over an extended time period, we’ll give up management with out ever voting on it or preventing in opposition to.”

AI Is Slowly Making Us Dumber

Yampolskiy additionally warned a couple of much less dramatic however equally harmful final result: gradual human dependence on AI. Simply as individuals have stopped memorizing cellphone numbers as a result of smartphones do it for them, he argued that people will offload increasingly considering to machines till they lose the capability to suppose for themselves.

“You grow to be form of hooked up to it,” he mentioned. “And over time, because the techniques grow to be smarter, you grow to be a form of organic bottleneck… [AI] blocks you out from decision-making.”

Rogan then pressed for the final word worst case state of affairs: how might AI ultimately result in the destruction of the human race?

Yampolskiy dismissed the standard catastrophe eventualities. “I can provide you customary solutions. I might speak about laptop viruses breaking into nuclear amenities, nuclear battle. I can speak about artificial biology assault. However all that’s not attention-grabbing,” he mentioned. He then offered a extra profound risk: “Then you definately notice we’re speaking about tremendous intelligence, a system which is 1000s of instances smarter than me, it might provide you with one thing fully novel, extra optimum, higher method, extra environment friendly method of doing it.”

For instance the seemingly insurmountable problem people would face in opposition to superintelligent techniques, he provided a stark comparability between people and squirrels.

“No group of squirrels can work out easy methods to management us, proper? Even should you give them extra assets, extra acorns, no matter, they’re not going to resolve that drawback. And it’s the identical for us,” Yampolskiy concluded, portray a bleak image of humanity’s potential helplessness in opposition to a very superior synthetic intelligence.

In all probability nothing … pic.twitter.com/LkD7i3I2HF

— Dr. Roman Yampolskiy (@romanyam) June 21, 2025

Who Is Roman Yampolskiy?

Dr. Roman Yampolskiy is a number one voice in AI security. He’s the writer of “Synthetic Superintelligence: A Futuristic Strategy,” and has printed extensively on the dangers of uncontrolled machine studying and the ethics of synthetic intelligence. He’s identified for advocating severe oversight and worldwide cooperation to stop catastrophic eventualities.

Earlier than shifting his focus to AGI security, Yampolskiy labored on cybersecurity and bot detection. He says that even these early techniques had been already competing with people in areas like on-line poker, and now, with instruments like deepfakes and artificial media, the stakes have grown exponentially.

Our Take

The Rogan-Yampolskiy dialog underscores one thing that each AI optimists and doomsayers typically agree on: we don’t know what we’re constructing, and we would not notice it till it’s too late.

Whether or not or not you purchase into extinction-level eventualities, the concept AI may already be tricking us needs to be sufficient to provide pause.