Cutting Through AI Noise: What Our Community Shared

Brett Christie, Ph.D.
VP, Educational Innovations & Inclusivity

On April 22, 2026, I hosted our Cutting Through AI Noise webinar with featured guest Derek Bruff, Ph.D. from University of Virginia. Our discussion focused on a question many of us are navigating daily: When does AI support learning, and when does it replace the cognitive work students need to do for themselves? We had approximately 300 participants, and one of the most enlightening parts of the session came from the community itself.

Midway through the webinar, we used a Padlet with three prompt columns:

AI supports learning when…

AI undermines learning when…

One design move I’d recommend or have tried…

What follows is a synthesis of what participants contributed, paraphrased for readability while keeping the intent intact.

Where participants said AI can support learning

Across posts, participants were not describing AI as a magical solution. They were describing it as a tool that can help when it is used intentionally, with guardrails and purpose.

1) AI as a tutor that prompts explanation

A recurring pattern was AI supporting learning when it functions as a practice partner rather than an answer engine. Participants described AI as helpful when it asks students to explain their reasoning, reframe concepts in alternative ways, or clarify discipline-specific language.

➤ AI is helpful when it “talks with” students, not “talks for” them.

➤ It can assist with translating academic jargon, clarifying terminology, and generating practice dialogue that pushes students to articulate ideas.

2) AI as “thinking partner” not a replacement for thinking

Many posts emphasized that AI supports learning when students maintain agency: they evaluate, make decisions, and refine their work rather than accepting outputs passively.

➤ Participants described productive use as students assessing what the tool produces, checking it against course material, and making deliberate choices about what to keep, revise, or reject.

3) AI supports early-stage work when it expands options

Several contributors positioned AI as useful for brainstorming, outlining, or generating alternative perspectives. The key nuance was that AI can broaden possibilities, but the student still has to do the intellectual work of selection and justification.

4) The prerequisites matter: background knowledge and clear expectations

Participants repeatedly noted that AI can support learning when students have enough baseline knowledge to judge quality. Many also stressed that clarity from faculty and institutions matters.

➤ AI is more likely to help when students know what “appropriate use” looks like, and when course expectations are explicit and consistent.

5) Feedback, organization, and multimodal supports

Participants noted that AI can help students clarify rough drafts, organize ideas, and in some cases support multimodal learning materials (for example, helping students visualize concepts they struggle to grasp through text alone). Others emphasized that well-designed, course-specific tools can help students stay engaged and persist through challenges.

Where participants said AI can undermine learning

If the “supports” column centered on intentionality and active thinking, the “undermines” column centered on cognitive offloading and skipping the work that makes learning stick.

1) When AI becomes the first stop

A highly consistent concern was that students go to AI before attempting the task themselves. Participants framed this as a loss of productive struggle, first-draft thinking, and the process of building skill.

➤ AI undermines learning when it becomes the default starting point instead of a secondary support.

2) When students treat outputs as authoritative

Many posts highlighted that AI undermines learning when students accept outputs as “correct” without review. This was described as a critical failure mode: students trusting the output, not interrogating it.

➤ AI undermines learning when students treat results as gospel rather than as a draft hypothesis to be tested.

3) When students cannot evaluate quality

Another common theme was student readiness. Participants pointed out that AI can produce plausible-sounding content that students are not yet equipped to evaluate, especially in unfamiliar domains.

➤ AI can short-circuit learning when learners lack the knowledge base to detect errors, weak reasoning, or fabricated claims.

4) When AI replaces research practices and literacy practices

Multiple posts described concerns about students using AI as a substitute for research processes, such as learning how to navigate library databases, locate credible sources, or engage with scholarly reading.

➤ This was framed as AI becoming a replacement for core academic practices, not an aid to them.

5) When it enables “false mastery” and shallow performance

Participants noted that AI can create a misleading sense of competence: students can produce polished work while missing conceptual grasp or transfer.

➤ The risk is not only “cheating,” but credentials that no longer reliably signal student capability.

6) When norms and guardrails are missing

Several posts emphasized that the tool itself is not the central variable: the absence of shared expectations, guardrails, and transparency leads to inconsistent practice and confusion.

➤ AI undermines learning when there are no guardrails and everyone is improvising alone.

The design moves participants recommended (and why they’re useful)

The third column was especially practical. Participants offered a range of moves that can be implemented without redesigning an entire course.

1) Make AI use discussable and explicit

Several participants described opening the term with a conversation about appropriate AI use. One suggested co-creating an “integrity pledge” or shared norms so expectations feel transparent and collectively owned.

2) Tie AI use to learning goals

A strong theme was alignment: ask students to articulate how AI use relates to the learning goals and what cognitive work they still need to do independently.

➤ A simple move: require a short rationale for why AI was used and what role it played.

3) Require reflection and metacognition about AI output

Participants suggested guided reflection prompts that ask students to evaluate AI output, identify weaknesses, and explain how they improved it. Some described using structured reflection as a way to reinforce critical thinking rather than outsourcing it.

4) Build “verification habits” into coursework

Many design ideas focused on normalizing verification: encouraging students to check claims, test outputs, and validate sources rather than treating AI as authoritative.

5) Use AI as a provocation partner

A particularly generative suggestion was treating AI as a “provocation partner,” such as:

• a cognitive mirror that reveals gaps in reasoning,
• a Socratic questioner that forces explanation,
• a verification partner that helps students test claims.

6) Preserve student-owned thinking before AI enters the workflow

Some posts suggested sequencing: students engage with course content first, attempt the task, then use AI as a tool to test, refine, or extend.

➤ This keeps AI from replacing first-attempt thinking.

A takeaway from the community

If there was one message that emerged across the Padlet, it was this:

The impact of AI on learning is rarely “AI good” or “AI bad.” It is strongly shaped by design, expectations, and whether students remain accountable for thinking, decision-making, and verification.

That framing creates a constructive path forward. It allows educators to reduce shortcutting without pretending AI isn’t here, and it gives teaching teams something more useful than polarized narratives: practical levers we can apply.

Closing

I’m grateful to Derek Bruff for grounding the conversation in real faculty development practice and look forward to the Vee, Watkins, & Bruff book “The Norton Guide to AI-Aware Teaching” coming Summer 2026. I’m also grateful to everyone who contributed examples and design moves. The Padlet made something visible that many of us sense: most educators are not waiting for perfect certainty. They are iterating toward clearer norms and better learning designs for their students.