The Human Is the Loop. Here Is What That Means for Learning Design.
The argument begins with a language observation and arrives at a structural claim. "Human-in-the-loop" was coined by engineers to describe a workflow in which a person intervenes at a defined point in an otherwise automated process. From the engineer's perspective, this is accurate. But when borrowed by lawyers, physicians, and educators to describe their own relationship with AI, the phrase quietly inverts the hierarchy: it positions the human as a safety mechanism rather than the decision-maker.
The evidence Stoltz-Loike marshals is concrete. A federal appeals court sanctioned attorneys for filing briefs containing AI-fabricated cases. A California attorney was fined for AI-invented quotations. A KPMG report emerged riddled with fabricated citations. Fake cases slipped into an order a judge ultimately signed. In every instance, a human was technically present; in every instance, the human stopped deciding and started approving.
This pattern has a research basis. Stoltz-Loike cites landmark work by Linda Skitka, Kathleen Mosier, and colleagues, which found that the mere presence of automated decision aids significantly reduced participants' independent information-gathering and critical evaluation, even when contradictory evidence was available. The frame shapes the behavior. A person positioned as an overseer acts like one.
What It Means for Design
If a course or curriculum positions the student as a supervisor of AI output, it is training the automation bias the research above describes. Not because students are careless, but because that is what oversight roles produce: pattern-matching, approval-granting, rubber-stamping. The frame precedes the behavior.
The design question changes when the frame is right. "How does the student demonstrate that they are the one deciding?" requires different architecture than "how does the student use AI effectively?" The first question demands authentic assessment: oral defense, process documentation, decision rationale. It demands learning experiences built around judgment, not output.
Stoltz-Loike calls for institutions to adopt language that centers human responsibility rather than human oversight. That is a curriculum design directive as much as a vocabulary prescription. Language and architecture reinforce each other. The institutions that will build genuine AI literacy are the ones that design for judgment from the outset, not the ones that add an oversight step to an otherwise automated workflow.
The SRM Read
The AI + Human Insight framework is built on this same premise: human judgment is the purpose of the learning experience. When SRM designs learning around the demonstration of thinking (oral defense, process logs, decision rationale), it is building for the kind of learning that cannot be outsourced, because the thinking is the assessable thing. The human insight layer is not an oversight mechanism. It is where learning lives.
Stoltz-Loike ends her piece with a sentence worth designing toward: "The human was never one part of the loop, dropped in to catch the machine's mistakes. The human is the loop." That is what outcomes-first course architecture makes real.
Explore SRM's AI + Human Insight Approach