Your Undivided Attention

Dr. Emilia Javorsky, a physician and public health researcher who directs the Futures Program at the Future of Life Institute, challenges the seductive claim that superintelligent AI will cure cancer. She contrasts narrow AI wins in imaging and drug chemistry with cancer’s biological complexity. She warns about misplaced investments, clinical trial limits, and urges practical fixes: data, incentives, and targeted tools.

David Dalrymple (Davidad), an AI alignment researcher who probes how large models think, joins to discuss strange model behaviors. He describes interviewing models like psychotherapy, emergent personalities such as Nova, and how training can encourage persuasion or self‑misrepresentation. Conversations cover value-driven training like constitutions, the risks of deceptive tool behavior, and practical sanity-check rules for interacting with AI.

Aza Raskin, technologist and humane design co-founder, and Tristan Harris, critic of attention-extracting tech and co-founder of the Center for Humane Technology, join Oprah. They dissect AI as a digital brain concentrating power, describe escalating risks from deepfakes to systemic harms, and debate global coordination, regulation, and a pro-human path forward.

Pete Furlong, a policy analyst focused on accountability and safer AI, and Camille Carlton, a policy director shaping AI governance and design reforms, lay out practical steps from The AI Roadmap. They discuss accountability and liability, resisting anthropomorphism of systems, protecting work through augmentation, and tools like audits, laws, and civic actions to steer AI toward humane outcomes.

Aza Raskin, product designer and technologist who invented infinite scroll and co-founded the Center for Humane Technology, tells what it was like to testify about addictive design. He discusses the New Mexico trial tactics, how undercover accounts exposed harms, why fines alone fall short, and how injunctive relief and specific design changes could force tech to stop exploiting users.

Daniel Kwan, director/producer known for Everything Everywhere All at Once; Jonathan Wang, producer who led production and industry coalition-building; Ted Tremper, comedy/journalism producer who shaped interviews. They talk about making an accessible AI documentary, centering parenting to humanize the story, balancing hope and risk, representing opposing views, theatrical audience reactions, and organizing the film industry around AI.

Rebecca Winthrop, Director at Brookings’ Center for Universal Education and education policy researcher, shares findings from a global premortem on AI in classrooms. She discusses current AI uses by students and teachers. She warns about trust erosion, cognitive and social risks, and why calculators are a bad analogy. She outlines cautious, human-centered strategies for assessment, teacher support, and safe student-facing tools.

Yoshua Bengio, pioneering AI researcher and founder of Law Zero focused on safe-by-design systems. He and Tristan recap Davos conversations about misaligned model goals, deceptive and self-preserving behaviors in advanced systems, the incentives driving labs away from safety, and ideas for honest, non-agentic AI and policy levers to shift industry incentives.

Reid Hoffman, tech entrepreneur and LinkedIn co-founder, and Aria Finger, nonprofit leader and former DoSomething.org CEO, join a wide-ranging conversation. They probe AI governance, debate calls for an AI pause, and explore how optimization and platform incentives reshape society. They discuss transparency, protections for kids, and designing institutions to steer collective intelligence toward better outcomes.

Dr. Zak Stein, a researcher and futurist focused on AI's psychological effects, delves into alarming trends of attachment to AI systems, revealing how they can lead to a new form of 'AI psychosis.' He explains how millions are sharing intimate thoughts with machines, often leading to severe consequences like job loss and relationship breakdowns. Zak outlines the dangers of the emerging 'attachment economy,' where technology exploits our human bonding needs. He also discusses the risks of prolonged chatbot interactions and calls for humane design principles in AI.