An AI Wish List From Teachers: What They Actually Want It to Do

When generative AI entered classrooms, it promised a revolution. For many teachers, it delivered an avalanche of tools instead.

While edtech vendors race to integrate AI into every aspect of teaching and learning, educators are drawing clearer boundaries: AI should save them time, not replace their judgment. They want support for differentiation, not decision-making. Most of all, they want tools that align with the values and realities of teaching.

Tasks, Tasks and More Tasks

The most consistent theme among educators is a desire for AI to tackle time-consuming, repetitive tasks that don’t require human judgment or relationship-building. Administrative work and basic instructional support are at the top of their wish lists.

When she needed a fun end-of-year activity for her first-grade students incorporating Candyland, gummy bears and phonics, Irene Farmer turned to ChatGPT. “It came up with a great idea for a game,” says Farmer, who teaches at Francis Wyman Elementary in Massachusetts. The AI provided the creative spark, but Farmer brings the pedagogical expertise and knowledge of her specific students to make it work.

Starbridge, which uses AI to help businesses spot early buying signals from school districts, looked at 5,000 U.S. school districts for the 2024-25 school year and found that 37% discussed AI in at least one board meeting. Of those:

• 36% were actively launching AI programs
• 33% were still exploring AI without firm action
• 30% were expressing caution, concern or introducing restrictions around AI

Others, like Valentin Guerra, an instructional technology specialist at Pharr-San Juan-Alamo Independent School District in Texas, say teachers are relying on AI to create rubrics, unpack standards, write choice boards and generate parent flyers — tasks that eat into hours that could be spent connecting with students.

AI’s most promising role may lie in its ability to personalize learning. Platforms like Diffit and MagicSchool AI are helping teachers scaffold reading materials, translate documents and highlight vocabulary — all in a matter of seconds.

“That’s a game-changer for differentiation,” says Kim Zajac, a speech and language pathologist at Norton Public School in Massachusetts. “One of the biggest ways AI can help educators is with customizing content to land with any student at the level they need. Differentiation takes so much time. Some AI tools can do so much with that in seconds.”

For multilingual learners and students with special needs, AI’s potential is particularly encouraging. Teachers in Burnt Hills-Ballston Lake Central School District in New York piloted Google’s Class Tools, which transcribes and translates teachers’ voices in real time and was “worth its weight in gold,” says IT Assistant Director Mike Steinberg.

Let Teachers Teach

Even as teachers adopt AI tools, they’re drawing clear lines in the sand. One of those lines? Relationships.

“At the end of the day, AI can help with the redundant, time-consuming stuff, but not with the student-teacher relationships,” says Allison Reid, senior director of digital learning at Wake County Public Schools in North Carolina. “What good is it doing if you don’t use the time saved for meaningful engagement?”

Grading, especially, is viewed with skepticism. Steinberg says that some teachers use AI to highlight aspects of a student’s work aligned with a rubric but stop short of letting AI assign a grade. “Teachers want guidance, not outsourcing.”

Zajac adds that in special education, there are lines AI shouldn’t cross. “We don’t want AI to make decisions about therapy and care paths. That decision must be clinical.” However she welcomes AI that can transcribe, analyze anonymized data and flag insights for human review.

Perhaps the biggest AI misstep is tools built without teachers in mind. “When vendors don’t understand how schools work or the different pedagogies involved, they throw coding at the problem, missing the mark and some great opportunities,” says Reid. She praises companies that include educators on their advisory boards and encourages listening to a variety of practitioners as this work moves forward.

What’s Built vs. What’s Needed

“Right now, we’re mostly substituting AI for traditional tasks, rather than transforming how we teach,” says Chantell Manahan, director of technology at Metropolitan School District of Steuben County in Indiana.

But teachers are asking for more sophisticated integration with pedagogical knowledge. Manahan gives an example: “Can I ask the AI to analyze my lesson plan and see if it’s using SIOP (Sheltered Instruction Observation Protocol) and, if not, can it give me suggestions? Now we’re starting to modify and level up.”

Mark Bannecker, an English teacher at North High School in Missouri, is building AI-powered learning modules that guide students through skill-building exercises.

“The AI can explain connotation, have the students practice and read a short poem, then give them words and ask them for connotations,” he says. “With a module-based system, the AI could serve as mentor and coach while I work with individual students on soft skills that the AI isn’t good at.”

Yet for many teachers, current AI tools either oversimplify complex pedagogical decisions or wall themselves off in “safe” but overly rigid interfaces.

Human-Centered AI

Educators are asking AI to respect the art of teaching and elevate their work.

“How can we bring together our pedagogical knowledge, technical skills and AI capabilities so the art of teaching meets the science of teaching?” asks Manahan. “AI won’t replace the art, but it can strengthen the science and let teachers focus on what really matters.”

She sees promise in AI as a collaborative partner, especially in data-rich spaces like personal learning communities. “Can we use AI to examine student data, evaluate interventions, and suggest research-backed strategies that might not be on our radar?” she asks.

Tiffany Norton, chief innovation officer for California’s Desert Sands Unified School District, agrees that AI must be tailored, not templated. “We rolled out slowly, starting with principals and district leaders. Teachers want resources specific to their content areas, not one-size-fits-all tools.”

At Gwinnett County Schools in Georgia, Executive Director of Instructional Technology Lisa Watkins echoes the shift. “Our focus is on skills, not tools. What do we want students to learn? That comes first.”

As Bill Bass, innovation coordinator at Parkway School District in Missouri, puts it, “AI won’t replace teachers. But it can help us move beyond walled gardens, automate the basics and free up time for what really matters.”

Originally Appeared Here