A report published this month by the American Psychological Association contains a finding that should reframe how every teacher evaluates educational technology: student engagement with an app is not evidence that the student is learning. The report, written by a multidisciplinary panel of psychologists reviewing evidence across digital learning tools for students ages 5 to 18, found that clicking, responding to animations, collecting rewards, and completing activities within an app are frequently mistaken for learning by parents, teachers, and administrators. The actual test of learning, the report argues, is what happens after the app is closed: can the student explain the concept, solve a new problem, or apply the skill in a different context?

For writing teachers, this distinction is particularly important because writing is a domain where the gap between performance and learning is easy to miss. A student using a grammar practice app may score 95 percent on comma exercises but still produce comma splices in their own essays. A student using an AI tool to improve their draft may submit a polished essay but be unable to explain their argument when asked. The app engagement data says one thing. The student's independent writing ability says another. The APA report frames this as a fundamental measurement problem: the metrics that apps report (time on task, completion rates, accuracy on exercises) measure platform engagement, not transferable learning.

The report also addresses generative AI specifically, noting that AI tools pose particular risks by improving students' immediate performance without building their underlying knowledge or skills. This echoes findings from earlier international research that documented students producing better essays with AI assistance while retaining less of the content afterward. For writing instruction, the implication is clear: a tool that makes the essay better is not the same as a tool that makes the writer better. The two look identical in the gradebook, but they produce very different outcomes over time.

The APA's lead psychologist for education summarized the practical takeaway in blunt terms: close the app, ask the student to explain, ask the student to solve a new problem or apply the skill in a different setting. The success you see on the platform might just mean the student is skilled at navigating that platform. This is not a call to abandon technology. It is a call to evaluate technology by the right criteria: not whether students are engaged, but whether they can demonstrate the skill independently after the engagement ends.

What This Means for Writing Classrooms

Writing teachers can apply the APA's framework by asking one question about any technology tool in their classroom: does this tool build the student's independent writing ability, or does it perform a writing task that the student should be learning to do themselves? The answer determines whether the tool is educational or merely performative.

• Grammar practice apps: useful for building pattern recognition, but only if the student can apply the skill in their own writing. Test by assigning an essay after the practice and checking whether the targeted skill transfers.
• AI writing assistants used by students: risky if the AI generates text the student did not produce. The APA report warns that performance gains from AI may mask learning losses.
• AI grading tools used by teachers: beneficial when they handle the scoring and feedback drafting while the teacher retains decision-making authority. The student's learning comes from engaging with the feedback, not from the tool's output.
• Gamified writing platforms: evaluate by what happens after the game. If students earn high scores on the platform but cannot produce the same quality writing independently, the engagement is not translating to learning.
• The APA's simple test: after using any tool, close it and ask the student to demonstrate the skill without it. If they can, the tool is working. If they cannot, the tool is entertaining.

Choosing Tools That Build Writers, Not Scores

The APA report does not recommend banning screens or abandoning educational technology. It recommends higher standards for what counts as evidence that a tool works. For writing teachers evaluating tools, those standards translate to specific questions: Does the tool ask students to do the thinking, or does it do the thinking for them? Does the tool's output require the student to engage with feedback and revise, or does it produce a finished product? Is there independent evidence that students who use this tool write better when they are not using it?

AI grading tools that follow the human-in-the-loop model pass this test because they are designed for the teacher, not the student. The tool scores essays and drafts feedback. The teacher reviews and adjusts. The student receives detailed, rubric-aligned feedback and is expected to engage with it through revision. At every stage, the cognitive work of learning to write stays with the student. The tool absorbs the mechanical labor of grading so the teacher can provide more feedback, faster, to more students. That is a tool that builds writers, not one that substitutes for writing.