The nation’s largest purveyor of licensing exams for teachers now offers a gauge of whether teachers have the skills to use artificial intelligence—the latest sign that the technology continues to infiltrate all corners of K-12 education.
Teachers are on the front lines of widespread AI adoption and upheaval in the last three years, but so far get limited training—or even guidance—in how to use the technology ethically and effectively in their classrooms. ETS’ Futurenav Adapt AI test, released last week, is intended to help education leaders get a clearer picture of what educators need.
The test uses three modules to evaluate the ability to recognize and understand generative and large-language-model AI in an educational context; navigate the technology ethically; evaluate AI-based tools and programs; and use AI in the classroom. It typically takes less than 30 minutes.
Forty-six states already use ETS’ suite of Praxis tests to gauge teaching skills and subject-specific content knowledge for teacher certification. The AI test was not specifically developed for licensure or other high stakes uses, but it does position ETS to respond if states decide to include AI topics in teacher certification in the future.
“There’s this universal concern around appropriate use: for protecting students, making sure students are using it appropriately, making sure teachers know how to guide them, and that teachers themselves are using it appropriately,” said Vince Dean, an associate vice president at ETS who leads the Praxis program.
Most teachers already use AI, but lack meaningful guidance
Most teachers, around 8 in 10, say they use AI in the classroom already, according to the nonprofit Center for Democracy and Technology, for everything from lesson plans, scoring guides, and quizzes to spotting students’ AI-enabled cheating.
However, early studies have found AI-based tools often produce less substantive lessons than human instructors. Often teachers don’t know how to write prompts to refine automatically generated content. Similarly, AI simulations may provide unrealistic versions of actual students at different ability levels.
Nearly all districts now use at least some AI-enabled tools or programs, and a majority report they have given teachers training in the technology. But for the most part, teachers say they’re on their own to learn how to use and respond to AI effectively.
“There’s a lot of choose-your-own-adventure stuff going on there,” Dean said. “We were finding that because people were engaging with different types of chatbots or different products, there wasn’t even a great understanding of base terminology: What is an LLM? How does it work? What are things to be aware of as you start using them?”
For example, test-takers must decide how to use public and private student information in an AI-based data tool. One scenario asks the teacher to work with a chatbot to develop and implement a lesson plan, assessing the teacher’s skill in writing AI prompts. In the final part of the test, the teacher reflects on how they developed the lesson with a virtual instructional coach (also portrayed by AI).
That is not as far-fetched of an idea as it may sound: Several districts have already started using AI-enabled tools in instructional coaching, both to help analyze student data and provide feedback and to help teachers reflect on their teaching. The final section of the AI test gauges both teachers’ skills in working with AI and also asks them to reflect on how they feel about doing so.
But testing and professional development for teachers shouldn’t replace clear guidance and support from school and district leaders, said Akhil Verghese, founder of Krazimo, which helps districts and other groups implement AI systems. (He was not part of developing the AI-skills test.) Only two states—Ohio and Tennessee—currently require districts to have an AI policy.
“Teachers are such motivated people that quite often it’s them who find the [AI] tools that they take into the school,” Verghese said. “But even if a teacher is brilliant, ... I don’t think it’s legally smart for a school to rely on teachers to use AI in whatever way.”
Test scores via a dashboard of strengths, weaknesses, and potential equity concerns are available for individuals and groups of people within a district. Dean said administrators can use the dashboards to prioritize professional development or coaching for individuals or groups of people, or to plan timelines for AI implementation.
ETS piloted the assessment last year with 75 secondary English/language arts, math, history, and English-as-a-second-language teachers at all levels of teaching and AI experience, according to Kateryna Komarova, a spokesperson for the organization.
