Oct 8, 2026 · 31 min
David Gerard challenges the case for accepting AI harms
“There’s no such thing as Rogue AI” David Gerard
The discussion tests whether AI’s risks reflect unavoidable technological trade-offs or preventable failures in engineering, security, and corporate governance.
- 1David Gerard argues that “rogue AI” narratives shift responsibility from flawed systems and companies onto supposedly autonomous machines.
- 2The episode links OpenAI’s contradictory messaging, disclosure problems, and weak apparent economics to broader questions about its leadership and future.
- 3Gerard says existential-risk ideology can both amplify fear of advanced AI and provide a rationale for continuing its development.
Don't miss
David Gerard argues that “rogue AI” is a marketing excuse for failures in systems built and controlled by companies.
The brief
David Gerard rejects the idea that chatbots possess volition, arguing that calling a failed system a “rogue agent” excuses poor engineering much like blaming a collapsed bridge on a rogue bridge.
The episode challenges Sam Altman’s claim that society must accept some bad outcomes for AI’s benefits, asking who bears those costs and whether current benefits justify them.
Gerard treats OpenAI’s shifting positions, delayed incident disclosures, and agent-security claims as signs of weak accountability rather than evidence of inevitable AI trade-offs.
The conversation turns to existential-risk ideology, tracing how fears of AI destroying humanity can simultaneously inflate the AI bubble and justify continued development.
Gerard closes by questioning OpenAI’s economics and Sam Altman’s leadership, arguing that debt, weak revenue, and funding dependence could threaten the company’s future.
Books & mentions
Some links are affiliate links — PodLume may earn a commission if you buy.
