Big Tech has lost control of the narrative | Thomas Germain
Nov 19, 2026 · 28m
More Technology podcasts · More News podcasts
Summary
Isaac and BBC tech journalist Thomas Germain discuss the legal and economic risks of agentic AI, noting that liability for hallucinations remains undefined while companies face massive financial losses. They explore how easily AI search results can be manipulated by "AEO" and the ethical complexities surrounding potential AI sentience and consumer harm. The episode concludes with Germain’s view that the AI industry is in a bubble, driven by unsustainable reliance on other AI companies for revenue.
Topics discussed
Sponsorships: Google Workspace and Citizens Bank
Intro: AI hallucinations and the hot dog eater story
Agentic AI risks, legal liability, and the Air Canada case
Compute constraints and liability for AI provider vs user
Regulatory shifts, Section 230, and AI publisher status
AEO/SEO manipulation and the ease of poisoning AI search
Sycophancy, AI psychosis, and Character AI lawsuit
Sponsorships: Citizens, Google Workspace, and Okta
Future liability precedents and AI harm settlements
AI sentience, Anthropic's policies, and ethical considerations
The slavery argument and self-serving narratives on consciousness
Sponsorships: Google Workspace, FT, and Wayfair
Philosophical risks of creating sentient AI
Is the AI industry in a bubble? Financial sustainability
Outro and final sponsorships
More from The Tech Report
Listen ad-free on Castria