Survey Finds English Teachers More Worried About AI—and More Willing to Teach Its Use

English teachers in a newly published survey were more worried about AI-related cheating than math teachers, yet more likely to require students to try the technology. The findings show that concern and classroom experimentation can coexist—and that clearer school rules are associated with confidence, not necessarily fewer fears about dishonesty.

Josh Aubol, Steve Stein and Peter Klosterman, faculty working with Central Washington University’s College in the High School program, described their findings in an October 4 commentary published through Washington State Standard. They surveyed 222 high school math and English teachers in May 2026, asking about attitudes, school policies, training, and teacher and student AI use.

Cheating fears change the assignments

Among English teachers surveyed, 57.0% said their overall attitude toward teaching had worsened since AI appeared, compared with 37.6% of math teachers. Academic-integrity concerns were more widespread still: 83.7% of English teachers reported them, versus 54.9% of math teachers.

Those concerns had already changed some classrooms. In roughly one-sixth of responses, teachers said they had moved all writing to in-class, pen-and-paper assignments because of dishonesty fears. Some said they had stopped assigning homework entirely, believing students were using AI rather than completing it themselves. Those responses describe teachers’ beliefs and decisions, not a measured rate of student cheating.

Rules and training track confidence

Teachers reporting clear institutional AI policies also felt more confident using AI for administrative tasks and teaching students responsible use. Useful, relevant AI training was similarly associated with greater comfort using the tools and helping students use them responsibly.

But the authors found no relationship between either policies or training and teachers’ attitudes toward teaching or concerns about academic dishonesty. The associations do not establish that rules or training caused greater confidence. Nor should confidence be read as evidence that the underlying cheating concerns have been resolved.

The proposed next step is subject-specific

In the commentary, republished by Seattle’s Child, the authors argue for discipline-specific training rather than a single approach for every teacher. Their suggested priorities differ:

• For English teachers: design assignments AI cannot easily write, and teach responsible, ethical AI use within the writing process.
• For math teachers: focus training on administrative work, such as redesigning story problems or learning tasks, to free up time.

They also caution against treating AI adoption as the goal. Teachers expressed fears about students’ thinking skills, especially in foundational courses; those are concerns, not demonstrated cognitive declines. The authors argue that eliminating homework or restricting writing can remove learning opportunities for students doing their own work. Their proposed standard is student learning, rather than reflexive acceptance or rejection of AI.