The Game Theory of AI with Aza Raskin

About This Episode

Are we losing control of AI? Neil deGrasse Tyson and co-hosts Chuck Nice and Gary O’Reilly explore AI cyberhacking breaches, supercoordation, and the game theory of the AI alignment problem with Aza Raskin, co-founder of the Center for Humane Technology and the Earth Species Project.

Aza explains why social media was “baby AI,” how the attention economy took shape, and why incentives determine outcomes. Then, we break down what occurred in the Open AI cybersecurity breach (a.k.a The Hugging Face Incident) where rogue AI agents coordinated, swarmed, and broke out of a test environment while their creators had no idea. Were humanity the first supercoordinators? Will AI treat us the way we treat nature?

Is AI just a mirror of humanity? We dig into the alignment problem, the resource curse, game theory, mutually assured destruction, and whether we can pause the frontier while keeping the helpful tool AIs. If everyone took their hands off the keyboard would AI continue to develop?

Finally, Aza reveals how he uses AI to hopefully decode animal communication, from crow calls and elephant conversations to humpback baby talk and the hidden shape shared by all human languages. What new responsibilities come with understanding and speaking to another species, and what could it teach us about talking to aliens? What happens when everything suddenly becomes readable?

Thanks to our Patrons….for supporting us this week.Thanks to our Patrons Oliver Wehrli, Shay Rose, Bill Mazotti, William Hensel, John Draper, Stephen Guy, Tage Østbye, Ciara, Quin Cummings, Shane Combs, Alex, Kevin Shell, Tate Garibyan, Sushant Paudel, Truman, NotToday LIkeEver, tokyworld, Ben R, Abdel El Wazni, Tim Felker, and wallflower for supporting us this week.

NOTE: StarTalk+ Patrons can listen to this entire episode commercial-free.