“Cutting Through the AI Hype: Predictive vs. Decision-Making AI”
👉 For more insights, check out the full book and audio summary here 📖
📕 Transcript and audio of this episode is available here for free
Title: AI Snake Oil
Author: Arvind Narayanan & Sayash Kapoor
Category: Technology & Innovation, AI, Data & the Future, Decision-Making & Critical Thinking
What if the most hyped AI applications are the ones that work the least well? Sam and Sophie sit with that provocative idea, unpacking how 'AI Snake Oil' separates the real from the fake.
They walk through the book's core distinction between predictive AI, which finds patterns, and decision-making AI, which tries to make choices for us. Real-world examples like automated hiring and predictive policing show why the latter often fails, with feedback loops, proxy variables, and missing values. The episode also covers the generative AI boom, explaining why ChatGPT is a 'supercharged parrot' that's useful for some tasks but dangerous for others.
If you've ever felt overwhelmed by AI hype or wondered whether to trust an algorithm with big decisions, this episode gives you a practical checklist: ask what it's predicting, how accurate it really is, and who's accountable when it fails. The takeaway: AI is a tool, and we can demand evidence, transparency, and accountability.
AI Snake Oil by Arvind Narayanan and Sayash Kapoor. If you want the full written summary, the whole library is on 7minutebooks.com/app — over 6,000 titles, unlimited access from $2.99 a month, $9.99 a year, or $19.99 lifetime.
Chapters
00:00The Core Argument: Hype vs. Reality00:47Predictive vs. Decision-Making AI01:49Real-World Failures: Hiring and Policing02:40Feedback Loops and Proxy Variables04:25The Generative AI Boom05:07Takeaway: Asking Better Questions
Topics
Related episodes
