Guest Post: When Mystery Powders Meet AI -- A Chemistry Lesson That Actually Worked

A guest post by Wess Trabelsi

Wess Trabelsi is a Technology Integration Specialist with Ulster BOCES and a Curriculum Design Specialist with AI Literacy Partners. Over the last year, Wess has assisted K-12 educators in New York State in implementing AI-integrated activities and lesson plans into classrooms across multiple disciplines. This year, he has developed a collaborative experimentation cohort of educators interested in determining when, where, why, and how AI may be used in the educational experience in ways that support and develop critical thinking, and he writes an incredible Substack blog called AI ∩ K12 = Wess, which I highly recommend.

Today, he shares his experience bringing AI into an 8th grade science class via a collaboration with a classroom teacher. His lesson plan is downloadable for free via our Assessment Redesign Services page, and this post aims to tell the story of its implementation—along with reflections, insights, and advice with respect to meaningful AI use in the classroom.

To make this resource as practical as possible, we’ve added Wess’ full lesson plan to our growing library of free AI-based lessons. You’ll find it alongside our ELA packet on The Great Gatsby and soon-to-launch history and math lessons. Just head to the Assessment Redesign Services page and click Download under the Science section:

I hope you find Wess’ design approach as meaningful as I do, and that his thoughtful blend of creativity, practicality, and vision sparks new ideas for how you might explore AI in your own teaching practice.

---

Last winter, an 8th grade science teacher approached me with a simple request: she wanted to try something new by integrating AI into a lesson but had no specific idea in mind. What emerged from our collaboration was one of the most engaging and successful AI-integrated lessons I've facilitated to date.

The Hook

During our planning conversation, I learned something that immediately caught my attention: these 8th graders had never performed a hands-on chemical experiment. They'd never mixed compounds, observed reactions, or felt the immediate sensory feedback of chemistry in action. This presented a perfect opportunity.

One of my core beliefs about AI integration is that technology cannot be used to circumvent the cognitive engagement required for learning. Some state guidelines emphasize this principle, though the list of tasks that AI cannot perform is shrinking at an alarming rate. Rather than playing a losing game of finding "AI-proof" activities, I focus on designing experiences where students genuinely want to engage with the content themselves.

The solution became clear: give students mystery substances and let them know they'd be allowed to experiment with them. The intrinsic motivation of manipulating chemical compounds would be our hook.

The Setup: Day One Mystery

Each pair of students received a small basket containing:

• A vial with a tablespoon of white powder labeled "NaHCO₃"

• Another vial with white powder labeled "C₆H₈O₇"

• A small beaker labeled "H₂O"

• A gallon-size Zip-Loc bag

• A plastic straw (which we later realized was unnecessary and caused confusion)

We deliberately deflected their immediate questions: "What is this? What are we doing today? What's it for?" Instead, we gave them five minutes of unstructured time to discuss, hypothesize, and raise questions with their partners.

Then we introduced the challenge: they would conduct a chemistry experiment AND learn to use AI as a research tool. The ground rules were simple—use AI to ask any questions needed to fill out their worksheet and design an experiment using the provided compounds. All approved experimental protocols would be conducted the next day.

Teaching AI Literacy on the Fly

Since these students were novices with chatbots, I provided a quick lesson on prompt engineering and AI interaction. Ideally, this would have been a separate session, as our 42-minute period didn't allow enough independent work time for all students to complete their worksheets.

As students worked with AI, I circulated with reminders that put them in control:

• "Remember, you're the boss—the AI should talk to you exactly how you want it to"

• "You are in control of the conversation"

• "Are you happy with the length of its answers?"

• "Ask it to explain at your reading level"

The goal was preventing the AI from overwhelming them with 26-step procedures when they needed simple, age-appropriate explanations.

The Results: Engagement Through the Roof

Day one was a resounding success. One hundred percent of students stayed on task, gaining familiarity with the chatbot while engaging deeply with chemistry concepts—identifying compounds, predicting outcomes, exploring endothermic and exothermic reactions.

We pushed the actual experiment to day three to ensure every group could complete their protocol.

Day Three: Chemistry Comes Alive

On experiment day, students barely glanced at their devices. Their attention was completely captured by the substances and ziplock bags. The room buzzed with animated discussions as pairs prepared for unknown outcomes.

The experiment itself was magical. Zero disciplinary issues, lots of questions and laughter, and genuine amazement as students experienced the cold sensation of an endothermic reaction. "Wow, it's getting super cold!" helped them connect abstract concepts to sensory reality.

Every group completed their work, engagement remained sky-high, and I'm confident most students could recall this lesson eight months later.

The Four Pillars of AI Integration

This lesson exemplified what I call my four pillars for successful AI integration:

1. Cognitive Investment

The lesson design must ensure students cognitively wrestle with content themselves, with AI as a partner, not a shortcut. We achieved this by creating an activity students genuinely wanted to master. The intrinsic motivation to manipulate chemical compounds was our secret weapon.

2. Assessment Validity

AI-integrated activities need ways to assess student understanding, not the AI's capabilities. While the AI may have done some heavy lifting for these novice students, they all met our learning targets for understanding basic chemical reactions. For more experienced students, I would restrict AI access to specific lesson phases.

3. AI Literacy Development

Every AI integration should teach new skills or provide practice opportunities. Since this was these students' first supervised chatbot experience, I focused on praise and gentle guidance rather than formal assessment. For experienced students, I would analyze chat transcripts and include AI literacy in the rubric.

4. Human Connection

Whenever possible, avoid AI as a solitary tool that isolates students behind screens. AI use should foster human-to-human interaction. In our lesson, AI "lubricated" peer conversations by bringing information and data points to discuss. Rather than following prescribed procedures, students had agency to define successful experiments, leading to enthusiastic debates between groups.

Lessons Learned

This lesson succeeded because we chose a safe, innocuous reaction (sodium bicarbonate and citric acid) for first-time experimenters while maintaining the excitement of discovery. The approach could easily scale to more advanced chemical reactions.

For future iterations, I plan to:

• Conduct AI literacy training in a separate session

• Collect and analyze student chat transcripts

• Develop more sophisticated assessment rubrics

• Gather formal data on student understanding

The Bottom Line

Beyond its instructional value, this was simply a joy to teach. When students are intrinsically motivated to engage with content, AI becomes a powerful amplifier rather than a replacement for thinking. The key is designing experiences where students can clearly see why they need to understand the material themselves, not just have a machine process it for them.

The teacher is already planning to repeat this lesson, and I'm excited to refine our approach. Sometimes the best AI integration happens when the technology fades into the background and human curiosity takes center stage.