Teachers Cautious on AI for Student Assessments, but Experts See Promise With Safeguards
Educators increasingly want to know whether AI can help them grade tests faster and give students better feedback. Experts say it can-but only if teachers stay in control of the process.
AI shows genuine potential for scoring assessments quickly and flagging student strengths and weaknesses. The technology can generate test questions, grade them, and provide feedback to teachers in a fraction of the time manual grading takes. But without proper oversight, the risks are real: privacy breaches, inaccurate results, and biased feedback.
Teacher enthusiasm lags far behind the technology's capabilities. In a 2024 EdWeek Research Center survey, 36% of educators said they think AI will make standardized testing worse within five years. Only 19% believed it might improve assessments. Even fewer are currently using AI to grade assignments.
The Human Must Remain in the Loop
Victor Lee, an associate professor of education at Stanford University, said the speed AI offers is valuable only if it doesn't introduce new problems. "Sometimes acceleration is great, but if it's accelerating with a lot of risks associated, then that's something we want to be really cautious about," he said.
Lee emphasized that teachers need to review all AI-generated feedback and add their own comments. This step prevents biases from embedding themselves in student evaluations and shows students their teacher actually read their work.
When providing assessment feedback, teachers communicate something beyond the score: "I paid attention to what you said, I see where your potential is." AI cannot replicate that message.
Building Teacher Competency First
Districts should require teachers to develop what Daniella McNamara, executive director of the Learning Engineering Institute at Arizona State University, calls "baseline literacy" on how AI works. Teachers need to understand what the technology can do, how it does it, and where it fails.
McNamara encouraged districts to let teachers experiment with AI tools. "The only way to really keep up is by using the tools with them," she said. "It's just like riding a bike. You have to ride the bike in order to learn how to do it."
Rather than asking AI to answer specific questions, teachers could learn to prompt it to explain events from multiple perspectives-a different political party's view, a historical figure's position-and build content from there. This approach deepens learning by exposing students to ideas through different angles.
Choosing the Right Tool for the Task
Lee pushed back against the idea that one AI tool should handle all assessment needs. Schools should ask: "What tool is appropriate for the task?" instead of trying to master every available option.
"There's the marketing hype of AI can do everything, and even if it could-which I would maintain it cannot-we don't want it to do everything," he said.
Before implementing any AI assessment tool, teachers should clarify their testing goal: Are they checking understanding of a specific topic? Gauging writing skills? Measuring reading comprehension progress? Different objectives require different scoring approaches.
Verification and Accountability
Humans must always verify AI's work. Teachers should check that scores and feedback align with established rubrics and catch errors before students see them.
"Because the AI doesn't know information about students contextually, leaving the AI to make all the decisions is not a good policy," Lee said.
When deciding whether to use AI for any assessment task, Lee offered a straightforward test: "Use AI for things with others that you would want them to use AI on with you."
Learn more about AI for Education or explore the AI Learning Path for Teachers to develop your own competency with these tools.
Your membership also unlocks:
