
Dreaming Against the Machine Episode 5: Metaphors for AI, with Shannon Vallor
May 12, 2026
Shannon Vallor, philosopher of technology and ethicist who wrote The AI Mirror, sketches how our metaphors shape AI thinking. She talks about AI as mirror, prosthetic, and accelerant. They critique Silicon Valley culture and libertarian tech myths. They also weigh narrow ML’s promise for public goods and the role of history in smarter AI policy.
AI Snips
Chapters
Books
Transcript
Episode notes
AI Is A Reflective Mirror Not A Mind
- The metaphor of AI as a mirror reframes models as reflections of historical human data rather than machine minds.
- Shannon Vallor explains mirrors can magnify, distort, and bring the past forward, causing AIs to present historical patterns as if they were the future.
AI Acts As An Accelerant On Existing Harms
- Treating AI as an accelerant means it intensifies existing social dynamics like bias and inequality when applied to those systems.
- Vallor uses the gasoline-can image: add AI to biased or unequal systems and you speed those harms up.
Audience Member Claimed Railroads Weren't Regulated
- Vallor recounts an audience member claiming railroads weren't regulated to argue regulation stifles innovation.
- She corrected him, explaining extensive railroad regulation actually enabled their widespread, coordinated growth.















