Fact-Checking w/ AI: Testing Claims Using Mike Caulfield’s New Critical Thinking with AI Approach

Описание: This is the second of two videos exploring different approaches to fact-checking a claim I heard on The Ezra Klein Show (“We Didn’t Ask for This Internet,” featuring Corey Doctorow and Tim Wu).

In the first video, I used Mike Caulfield’s SIFT framework. In this one, I experiment with his emerging work on how artificial intelligence can (and cannot) support fact-checking.

The claim under review:

Are gig-economy nursing platforms using credit scores and financial data in ways that contribute to wage discrimination or a “race to the bottom”?

In this video, I:

Compare Google’s AI Overview with deeper AI search modes
Reflect on how AI can amplify rabbit trails if our question isn’t well-formed (aka oops, I did it again 🤦‍♀️🫠)
Practice iterative prompting (instead of treating AI as a one-and-done answer machine)
Use Caulfield’s AI-informed approach, including Toulmin-style argumentation
Ask for evidence for and against the claim
Surface nuance around algorithmic pay-setting, contractor status, union protections, and intersectional impacts
Request academic sources and links and evaluate how well AI delivers

What emerges is more complex than a simple yes/no answer. Some gig nurses report higher wages and flexibility. At the same time, investigative reporting and peer-reviewed research raise concerns about opaque algorithms, contractor loopholes, and structural inequities.

This experiment reinforced something important:

AI can be powerful — but only when used dialogically, iteratively, and critically. It is not a shortcut around critical thinking.

If you teach media literacy, information literacy, digital literacy, or simply want to think more carefully about how to use AI in your own fact-checking process, I hope this “learning out loud” walkthrough is helpful.

📌 Watch the companion video for the SIFT-based approach.