How Teachers Can AI-Proof Assignments

A Gallup poll recently uncovered an interesting relationship between Americans and artificial intelligence. As AI use has increased, trust has not grown alongside it.

The percentage of those polled who use AI regularly jumped from 39% to 51% between 2025 and 2026. Those numbers are flipped, however, when you survey trust.

Of the same individuals polled, 28% in 2026 say they trust AI, compared with 32% in 2025.  

The report breaks down this erosion of trust:

“When asked what would increase people’s trust in AI, no single intervention generates widespread consensus, but three garner equally strong support: greater transparency from companies about how AI makes decisions (49%), testing of AI by independent experts (48%), and testing AI systems for fairness before they are deployed to the public (46%).”

Teachers, in particular, sit on that balancing beam of usefulness and trust, recognizing that AI competence matters but staying careful not to fall into reliance. For example, nearly half of educators use AI to boost efficiencies, however, most do not have formal guidance on how to best use these applications. This becomes a challenge when schools must consider student AI use, ultimately approaching the technology with curiosity rather than as a way to cut corners.

In the classroom, teachers have some control over AI use. There is the ability to monitor assignments, ask students to share their findings and limit technology use overall. However, this guidance ends once school wraps for the day, with students gaining access to the internet and AI tools to support homework.

The more a brain depends on technology to solve problems, the fewer critical thinking skills it develops, which is why engaging in schoolwork without AI interference is necessary.  Relying on AI, rather than using it as a tool, takes away the critical thinking needed to make problem-solving easier.

To help guide educators as the AI landscape evolves, the suggestions below outline ways to AI-proof assignments and foster critical thinking. 

Activities and assignments that support critical thinking

This summer, NPR polled educators and discovered 80% believe schools should teach responsible artificial intelligence use. This poll was published nearly a year after we published Modeling Responsible Artificial Intelligence (AI) in the Classroom, showing that AI is a daily part of our lives. Reflecting on its growth in just the past year, it is now apparent the technology isn't going away anytime soon.

Our new way of life means educators must take extra steps to prevent AI use on coursework and think critically about how to properly ‘AI-proof’ activities and assignments. These extra steps help preserve the critical-thinking skills students learn through classroom and homework assignments, something more than half of educators worry about.

"I care about [my students]. I want them to be able to look at the world and figure out things for themselves, not rely on a piece of software," Biology Teacher Michele Naber told NPR. "If we stop questioning what it says, we can be led to believe anything. And that's what really scares me."

Some activities and assignments are easier to tackle with AI than others, so educators should identify these categories and the activities that fit into each one. Below are suggestions for educators who want to engage students while avoiding AI use.
 

Handwritten homework assignments

In August, Lebanon Township in New Jersey announced its district would limit screen time to two hours per school day.

"You asked a student to open a notebook and take out a sheet of paper just to write something down on scrap, and they'll take out a sheet and put it backward, and upside down and start writing on it," Language Arts Teacher Julie Garvey told CBS News. 

Handwritten assignments can provide a simple way to move learning away from a screen while encouraging students to slow down and engage directly with the material. Teachers can ask students to complete short written responses, journal entries, reading reflections, problem-solving exercises or brainstorming activities by hand for any subject. 

Asking students to explain a concept in their own words, show their work or write a personal response also gives teachers a better window into how students are thinking, further improving connections.

In-class discussions

AI can provide an answer, but it can't replace the experience of explaining an idea to classmates, responding to an unexpected question or reconsidering an opinion after hearing another perspective.

Build regular discussion into lessons by asking open-ended questions that require students to explain why they think something rather than simply identify the correct answer. Socratic seminars, small-group discussions, think-pair-share activities and classroom debates can all encourage students to listen, respond and build on one another's ideas.

Teachers can also ask students to defend an interpretation, identify evidence that supports their position or explain why they changed their minds during a discussion. These activities make reasoning visible while helping students develop communication and critical-thinking skills.

Presentations and diorama-building

Give students opportunities to make something that cannot be reduced to a simple AI-generated response. Presentations, posters, models, dioramas, demonstrations and other hands-on projects require students to make decisions about how they will communicate what they have learned.

For example, instead of asking students to write a report about a historical event, have them create a model, visual timeline or presentation and explain the choices they made. A science student might build a physical model of an ecosystem, while an English student could create a visual representation of a novel's setting or themes.

These projects also give students opportunities to present their work to classmates. Explaining a project aloud can reveal whether a student truly understands the material or has simply reproduced information.

Require process, reflections and reasoning to be part of the assignment

Help students develop metacognition by making them hyperaware of their own learning journeys.

Rather than evaluating only the final product, ask students to document how they arrived at it. This could include brainstorming notes, research questions, drafts, revisions, problem-solving steps, source evaluations or a short reflection explaining what they would change.

Reflection questions can be simple: What was the most difficult part of this assignment? What strategy did you use when you got stuck? Why did you make this choice? What did you learn from revising your work?

When the process becomes part of the assignment, students have more reason to engage with the thinking behind their work. It also gives teachers additional evidence of learning beyond a polished final answer.

Incorporate ambiguity

What's the answer if there is no right answer? 

Adding ambiguity to work, discussions and lessons helps students learn that not everything is black and white. Multiple routes can lead to an answer, and sometimes there is no answer at all.

Teachers can incorporate questions with competing interpretations, incomplete information or multiple defensible solutions. In English class, students might examine whether a character's actions were justified. In social studies, they might consider how different groups experienced the same historical event. In math, they could find multiple ways to solve a problem and compare the advantages of each approach.

Ambiguous assignments make it more difficult to simply ask AI for "the answer." More importantly, they teach students that learning often involves uncertainty, judgment and interpretation.
 

Highlight the importance of turning to experts first

Expert voices will always be more authoritative than AI, so teach students how to identify experts and decipher their research and findings.

Students should understand that AI can summarize information, but it doesn't automatically make that information reliable. Teach them how to identify qualified researchers, academic institutions, government agencies, professional organizations and other credible sources relevant to their assignments.

Make source evaluation part of the learning process. Instead of simply asking students to "research" a topic, ask them to identify who produced each source, what evidence it provides, how that evidence was gathered and whether other experts agree.

This approach turns research into more than finding information. Students learn how knowledge is created, evaluated and challenged—and why going directly to credible human expertise remains an important part of learning in an AI-driven world.

If you’re considering following your dream of teaching, Rutgers Alternate Route can offer you the support and training you need to succeed. Be sure to follow Rutgers Alternate Route on Twitter or sign up for Alternate Route’s monthly newsletter to receive more information and stories from the field of education.