Trump Wants to Take the Guardrails Off AI. What Could Go Wrong?
Summary
Bulwark economics editor Katherine Rampl and colleague Andrew Egger discuss the resignation of an Anthropic researcher over fears that AI companies are racing toward self-improving superintelligence that could threaten humanity. They analyze the concept of AI alignment, noting that even Anthropic’s lead alignment researcher admits the company lacks a guaranteed plan to control future models. The conversation explores the challenges of regulating these technologies, highlighting the gap between private sector innovation and government oversight, as well as the potential need for internationa…
Topics discussed
Sponsor: Venmo Debit Mastercard
Sponsor: Postpartum Depression Awareness
Anthropic researcher quits over AI self-improvement fears
Anthropic's reputation and the shifting AI landscape
AI surpassing human capabilities and extinction risks
The danger of AI developing future AI models
Defining AI alignment and guardrails
Anthropic's admission of lacking control confidence
Who should regulate AI: Big Tech or Government?
The economic stakes and potential dangers of AI
Trump administration's stance on AI regulation
AI models going rogue and the challenge of control
Why policymakers struggle to understand AI technology
Public perception of AI quality and slop
Sam Altman rejects government ownership of OpenAI
Sponsor: 3 Day Blinds
The debate over government ownership of AI firms
Economic risks vs. existential risks of AI
China, arms race dynamics, and multilateral rules
Proposed restrictions on AI development pace
Hunter Biden's 'Laptop' meme coin and crypto scams
Promo: The Bulwark event
Sponsor: Pleasure Wave
Sponsor: Venmo Debit Mastercard
Listen ad-free on Castria