Expert: When it comes to AI in the classroom, focus on evidence of learning — not student use

BUFFALO, N.Y. — As generative AI becomes commonplace in schools and workplaces, educators question what kids actually know when AI is doing much of the work for them — especially when it comes to writing.

Some educators are returning to handwritten assignments, blue books and in-class essays to ensure students demonstrate their own skills. And while there may be times when students need to work without AI so educators can assess their unaided performance, banning it from the classroom doesn’t address the larger challenge of preparing students to use the technology effectively and responsibly.

Sam Abramovich, associate dean of academic affairs for the Graduate School of Education at the University at Buffalo, explains how educators can incorporate AI without compromising critical thinking and meaningful evaluation.

“Traditional assessment models are increasingly failing us in an AI world,” Abramovich says. “For decades, educators have treated a finished assignment as evidence that a student possesses the knowledge and skills reflected in it. Generative AI breaks that assumption. But the answer cannot simply be to ban the technology. We have to redesign assessment so that we can still make valid claims about what students know and can do.”

Abramovich researches and writes about how generative AI is changing assessment and learning and what educators can reasonably evaluate about a student’s own abilities in an AI-assisted world. Abramovich teaches courses in which students are expected to use AI and answer the questions that arise when technology becomes part of the learning process rather than something students are prohibited from using.

Among the issues Abramovich can address are how:

• AI changes student assessment.
• Educators can determine what students actually know when AI contributes to their work.
• Schools can prepare students to use AI without allowing it to replace the thinking students need to develop.
• Educators can incorporate AI without surrendering learning goals.
• Higher education can best prepare students.

Abramovich can translate these rapidly evolving questions for educators, policymakers and the general public, helping move the conversation beyond whether students should simply be allowed or prohibited from using AI in the classroom.