A survey of 303 Wisconsin educators and 132 school professionals nationwide, conducted from spring 2025 to spring 2026, found that while academic dishonesty dominates concerns about generative AI in schools, a more fundamental problem has emerged: teachers can no longer tell what students actually understand when polished work can be generated in seconds. An estimated 84% of high school students used generative AI for schoolwork in 2025, according to College Board.
In the Wisconsin sample, roughly 65% of respondents identified cheating and plagiarism as a concern, compared with 74% nationally. But nearly half pointed to something deeper. In Wisconsin, 47% of respondents said "difficulty in assessing student learning when AI is used" worried them. That figure rose to 53% in the national sample.
What the survey revealed about student reliance
When asked about AI's impact on student behavior, 29% of Wisconsin respondents and 40% of national respondents selected "increased student reliance on AI." Another 19% and 33%, respectively, chose "reduced critical thinking or problem-solving." These numbers suggest educators are watching students outsource cognitive work that assignments were designed to build.
Teachers have long known that a finished assignment is imperfect evidence of learning. A parent might help too much. A student might copy from a friend. Generative AI makes that problem more visible and more complicated. A homework task like writing a paragraph explaining the theme of a short story now produces results that appear organized and accurate - but reveal nothing about whether the student read the story, identified the theme, or articulated it independently.
Many educators reported using AI themselves for planning, communication, documentation, differentiation, and administrative tasks. Their concerns were not about rejecting the technology outright. They were about preserving meaningful evidence of learning when AI can produce academic work on demand. For teachers seeking structured guidance, an AI Learning Path for Teachers addresses these classroom challenges directly, covering tools and strategies for assessment in an AI-enabled environment.
Detection tools create more problems than they solve
Some teachers have turned to AI-detection software. In a 2025 national survey of sixth- through 12th-grade public school teachers, 43% reported using these apps regularly, while another 27% had experimented with them. The tools, however, are unreliable in both directions.
One study of 14 AI-detection tools found false-positive rates as high as 50% and false-negative rates as high as 100%, depending on the tool. About 20% of AI-generated texts were misclassified as human-written. That rose to roughly 52% when AI-written text was manually edited and 71% when machine-paraphrased. Other researchers found that detectors falsely flagged nonnative English writing as AI-generated at an average rate of 61.3%.
How some teachers are adapting
Rather than abandon writing assignments, some educators are redesigning them to make student thinking visible. They ask students to show or explain their process, include oral components alongside written work, or write more in class. Paper-and-pencil tasks are returning when teachers need to see independent thinking. If the goal is writing fluency, teachers need to see students write. If the goal is reading comprehension, students need to explain, apply, or defend their thinking.
Researchers who developed the Artificial Intelligence Assessment Scale have argued that educators should identify what level of AI use makes sense based on the learning outcomes being measured. One assignment might require no AI because the teacher needs to see independent writing. Another might allow AI for brainstorming but require students to submit original notes and a final reflection. A third might ask students to critique an AI-generated answer and explain what is accurate, incomplete, or misleading. Broader resources on AI for Education offer frameworks for making these distinctions in practice.
Policy remains patchy. Only 33% of Wisconsin respondents and 29% of national respondents said their district had a formal AI policy. Teachers and students alike could benefit from clarity on how and when AI use is permitted.
Why this matters for educators
The survey points to a practical challenge that cannot be solved with detection software alone. The goal is not to catch every instance of AI misuse - that is likely impossible. The goal is to design learning tasks where teachers can still answer the question that matters most: what does this student actually understand. For educators, that means auditing assignments not for how easily AI can complete them, but for whether the task itself produces visible evidence of student thinking. If it does not, the assignment may need to change, not the plagiarism policy.
Your membership also unlocks:
