Was this forwarded to you by a friend? Sign up, and get your own copy of the news that matters sent to your inbox every week. Sign up for the On EdTech newsletter. Interested in additional analysis? Upgrade to the On EdTech+ newsletter.

The best thing I read all week was the order prohibiting fireworks in my state this coming July 4 holiday. Given the ongoing drought and fires, that is a relief. But what about EdTech?

In my reading this week, I found myself thinking that AI has taken two steps forward and one step back. We're asking better questions about learning and work, but we're still letting hype shape investment.

For most of the past three years, we've been asking the wrong question about AI and learning. We keep going back and forth about whether to ban it or embrace it and, if we do embrace it, how to use it in a course.

This is the argument made by a pair of engineering faculty members from the University of Zagreb, Croatia in a new article in which they make the case for what they term placement. After reviewing a broad range of recent research on the impact of AI, they come down in favor of its use—but only in certain places within the learning process.

The critical question, then, is how to decide where AI should be used in a learning experience so that it does all of those good things, rather than the kinds of things that can harm student learning, such as simply providing the answers.

To help instructors answer that question, the authors develop a six-stage framework for learning: Prime, Probe, Point, Attach, Strengthen, and Test.

They argue that AI belongs in some stages but not others. The first and last stages (Probe and Test) should remain AI-free, while AI can be used as scaffolding through the middle stages.