What Does “Investigate the Evidence” Mean?
Source: Mike Caulfield Substack
Author: Mike Caulfield
Published: 2026-03-31
Source type: guide
Original source: https://mikecaulfield.substack.com/p/what-does-investigate-the-evidence
Private backup: the full article text is archived in the private repository at archives/articles/mikecaulfield-substack-com-what-does-investigate-the-evidence.source.md. It is not published on the public Quartz site.
Summary
Mike Caulfield explains a proposed SIFT-for-AI move called “Investigate the evidence.” In his evolving AI adaptation of SIFT, source verification remains important but is no longer always the first move. Instead, users first get enough context into the AI, then ask evidence-focused follow-ups that gather, categorize, and source relevant context. Using a social media claim about Bowlero, bowling leagues, corporate ownership, and Robert Putnam’s Bowling Alone, Caulfield shows that an initial AI answer may restate a thesis while an evidence-focused follow-up produces a more nuanced pro/con view.
Pull quotes
Evidence-focused follow-up
“The second thing I found in my experimentation was that the most useful sort of follow-up for the things I teach was what I call an evidence-focused follow up.”
Organized evidence, not just answers
“Investigate the evidence. Ask evidence-focused follow-ups. Gather and categorize important context. Get organized and sourced information, not (just) answers.”
A little work to get smarter
“However SIFT evolves, the focus is still “How do we do a little bit of work to get a lot smarter?”; that part will never change.”
Big ideas
- Students need to check AI answers against real evidence
- Students need to bring the purpose; AI should not supply it for them
- AI literacy has to be taught inside real subjects
- AI tools should be judged by the work they will actually do
Claims
- AI-assisted inquiry should ground claims in evidence
- AI-assisted inquiry should ground claims in evidence
- AI literacy should teach students what to do with AI, not just what to think about it
- Prompting AI is a literacy practice, not just a technical trick
- Educators need sustainable ways to keep up with AI
Key evidence and examples
- Caulfield reports experimentation with more than a thousand tests, finding that neutral follow-ups generally improved AI Mode answers.
- The Bowlero example asks for evidence for and against the claim that Bowlero is bad for league play.
- The follow-up produces a more balanced framing about critics, supporters, profit priorities, and financial viability.
- Caulfield cautions that users should still track better sources, especially when citations are weak or Reddit-based.
Education relevance
Very high for AI literacy, media literacy, research instruction, civic reasoning, source evaluation, and classroom routines for using AI without outsourcing judgment.
Durability note
This page is durable as a concrete AI-literacy move: the exact AI Mode examples may age, but evidence-focused follow-up remains a teachable habit.