Shirtloads of Science

Deepfakes Online and The Future of Truth with Dr. Anna Broinowski (464)

I'm joined by Dr. Anna Broinowski - Senior Lecturer and Director of the Master of Film and Screen Arts at Sydney College of the Arts, University of Sydney - to unpack the rise of deepfakes and synthetic media. We discuss the emergence of deepfake pornography in 2017, the fact that around 90% of deepfake content targets women and girls, and the growing use of AI in propaganda. From political manipulation to fake ads impersonating me (!), this episode explores how deepfakes are eroding trust, and what that means for the future of media, filmmaking and truth more broadly.

The Rise In Nitrous Misuse with Assoc. Prof. Angela Chiew (463)

Nitrous oxide (better known as "laughing gas") is widely used in medicine as a safe anaesthetic and pain reliever. It's also found in whipped cream canisters and even used in car engines. But Angela Chiew, toxicologist and Associate Professor at UNSW, joins me to set the record straight: the growing recreational misuse of this easily accessible gas is causing serious long-term harm, including nerve damage, cognitive damage, weakening of the spinal cord and mobility impairments. Cases of misuse are rising globally, likely driven by how easy it is to legally purchase the drug in many countries. This episode unpacks the hidden dangers behind a gas many once thought was just a harmless high. Linkedin: Angela Chiew X: AngelaChiewA

Will AI Take Our Jobs? Part Three of the Great AI Safety Debrief with Dr. Petr Lebedev (462)

In the final episode of the Great AI Safety Debrief series with Dr. Petr Lebedev, we discuss job eradication, AI alignment and the rise in cases of AI deceiving humans. Petr unpacks ideas like AI scheming, situation awareness, parasitic relationships between humans and machines, and even the possibility of AI developing belief systems of its own. We end by zooming out to existential risk, Alan Turing's early warnings, and why aligning AI with human values may be one of the most difficult (and urgent) challenges humanity must face. Linkedin: Petr Lebedev palisaderesearch.org/

Is AI Outsmarting Us? Part Two of the Great AI Safety Debrief with Dr. Petr Lebedev (461)

In part two, Dr. Petr Lebedev and I dive deeper into how modern AI systems are grown, not programmed, and the emergence of Shutdown Resistance. We explore AI performance in math and coding competitions, the rapid evolution of ChatGPT models, and why companies are already letting AI write the majority of their code. Petr explains how huge investments globally are pushing these systems forward faster than we can fully understand or control. The final episode in this 3 part series drops next week! Linkedin: Petr Lebedev palisaderesearch.org/

The Great AI Safety Debrief Part One with Dr. Petr Lebedev (460)

Intelligence is emerging from sand and less than 1% of the population seems to care! Physicist and science communicator Dr. Petr Lebedev joins the podcast following his SXSW talk to explore the rapid advancement of AI and how it's dramatically shifting the global status quo. From completing a PhD in physics and writing for Veritasium, to leading science communication at Palisade Research, Petr steps in to bring some clarity to the weird, wild, and worrying sides of AI. Did you know that no one truly understands how large language models actually work? We go deep on misaligned systems, humanoid and dog robots, and why many experts believe AI could pose a serious global risk. Is AI our greatest tool? Or something we're slowly losing control of? This conversation is big, so we've split it into three parts. Part two drops next week.