As the debate about the role of artificial intelligence carries on, teachers once again find themselves at the forefront of a discussion about the impact of new technologies on education. For many teachers, the outlook isn’t great.

As one teacher wrote in a recent survey we conducted, “I don’t know whether teachers or students will be relevant in the future.”

Yet not all teachers’ attitudes toward AI are so bleak, and there may be some evidence as to what can help teachers use these tools in a way that centers student learning.

In May, as faculty members working with Central Washington University’s College in the High School program, we surveyed 222 high school math and English teachers. The survey asked teachers a variety of questions about their attitudes toward AI, institutional policies, professional development opportunities, their own AI use, and student AI use.

Teacher attitudes and AI

Teachers reported significant differences in their attitude toward teaching since AI has appeared depending on academic discipline, with 57% of English teachers reporting that their overall attitude toward teaching has worsened since AI appeared, while only 37.6% of math teachers reported a similar change in attitude.

The difference between the two disciplines was even larger when it came to concerns about academic integrity, with 83.7% of English teachers reporting academic integrity concerns, and only 54.9% of math teachers reporting the same concerns.

Yet despite a worsening attitude toward teaching and concerns about academic integrity, English teachers also reported greater confidence teaching students to use AI responsibly (45.5% for English and 24.4% for math) and were more likely to require students to experiment with AI in their classes (34.1% for English and 14.6% for math).

Although this increased confidence seems contradictory at first, it may reflect the type of AI tools teachers and students have at their disposal. Large language models, which produce convincing (if sometimes factually inaccurate) writing as their output, are often reported to make frequent mistakes when completing simple mathematical calculations.

Because of this, the difference between English and math may reflect the tools themselves, rather than any difference in preparedness.

Teacher confidence

In addition to differences in attitude, our analysis also revealed a relationship between institutional policies and teacher confidence in using AI and teaching students to use AI. Teachers who had clear institutional policies around AI use also reported feeling more confident using AI for their own administrative tasks and in teaching students to use AI in a responsible way. At schools without clearly laid out policies, teachers reported less confidence in these areas.

A similar relationship also appeared for teachers reporting useful and relevant professional development opportunities centered around AI. Teachers who received useful AI professional development also reported feeling more comfortable using AI themselves and helping their students to use AI responsibly.

Notably, our results did not show a relationship between institutional policies or professional development and attitude toward teaching or concerns about academic dishonesty. It appears that the effect of institutional policy and professional development on these factors may be limited.

Although it is perhaps not surprising that teachers who are trained on a new technology like AI feel more confident using it, it does serve as an important reminder that tailored professional development opportunities remain a vital tool in increasing teacher confidence around new technologies like AI.

In particular, the difference in attitude between disciplines makes the need for discipline-specific professional development opportunities clear. For English teachers, this may include professional development focused on designing assignments that cannot easily be written by AI and training students to use AI responsibly and ethically as part of the writing process. For math teachers, it might mean providing trainings centered on how AI can be used to free up time from administrative tasks like redesigning story problems or learning tasks.

Effective AI policies

Equally clear is the need for effective institutional level policies around both teacher and student AI use. Without clear institutional guidance on what type of AI use is permissible for teachers and students, teacher uncertainty around what is and is not acceptable AI use will continue.

It also remains important that institutional policies and professional development opportunities do not unthinkingly embrace AI technology without a consideration of how AI use might impact student learning. In their survey responses, many teachers expressed concern that student AI use might lead to cognitive declines in students, particularly when students are allowed to use AI in foundational courses that teach critical thinking.

Yet a considered approach remains necessary if we want teachers to make classroom-level decisions rooted in student learning.

In roughly one-sixth of survey responses, teachers noted moving all writing to pen-and-paper, in-class writing assignments due to fears over academic dishonesty. Some teachers even wrote that they have stopped assigning homework entirely due to their belief that students are no longer completing assigned work without the use of AI.

Although these fears may reflect the feeling (reported by some teachers) that most of their students have stopped doing their work and are just using AI to complete assignments, these types of classroom decisions arguably take away opportunities for learning from the students who are doing their assigned work and trying to learn.

If math teachers had simply decided to stop teaching arithmetic because of the broad release of graphing calculators in the 1990s, student cheating may have gone down – but a whole generation of students would have also been left unprepared to do simple math.

A considered approach

Instead of reflexively rejecting or wholeheartedly embracing AI in the classroom, clear institutional policies and high-quality discipline-specific training are both necessary to ensure we are making decisions rooted in student learning. We opened with the observation of one teacher who questioned whether teachers or students would remain relevant in the future. The findings of our study suggest that both remain indispensable.

AI may reshape teaching and learning in schools, but it does not eliminate the need for teacher judgment or student effort. Learning still depends on teachers making informed decisions about instruction and on students doing the intellectual work required to develop understanding.

Just as importantly, the best classrooms are built on relationships. Teachers encourage, challenge, and support students in ways that technology cannot.

As schools continue to navigate the opportunities and challenges presented by AI, the more important question may not be whether teachers and students remain relevant, but how AI can be used while preserving the human connections that inspire students to learn.

This story is republished from the Washington State Standard, a nonprofit, nonpartisan news outlet that provides original reporting, analysis and commentary on Washington state government and politics.

Josh Aubol is a senior lecturer at Central Washington University. He serves as faculty liaison for English and Humanities for the university's College in the High School program. 

Steve Stein serves as a faculty liaison for Central Washington University’s College in the High School program. He has also worked as a high school math teacher, an athletic director, and a school principal.

Peter Klosterman is an associate professor of mathematics at Central Washington University who enjoys investigating how students interact with mathematics.