Original

Jim Roche is the Lead Consultant and Founder of Mountainside AI Education Consulting, where he partners with forward-thinking independent and charter schools to transform "AI Chaos" into sustainable, academic mastery.

​

With over 20 years of classroom experience—including nearly a decade teaching Chemistry and Environmental Science at the prestigious Trevor Day School in New York City—Jim understands the delicate balance between rapid technological innovation and rigorous pedagogical standards. He is a hands-on educator who has navigated the challenges of modern classrooms first-hand, making his consulting highly practical, empathetic, and tailored to the unique culture of independent education.

​

Jim is the pioneer of the Generative Thinking Model (GTM), a groundbreaking framework designed to intentionally integrate AI into the curriculum. Rather than replacing student effort, GTM uses generative tools to actively augment critical thinking, deepen problem-solving capabilities, and foster a healthy, ethical AI-student relationship.

​

As a sought-after voice on the intersection of AI and education, Jim has presented his innovative strategies at Columbia University (Fall 2025) and the SITE Conference in Philadelphia (March 2026).

​

Through Mountainside AI, Jim helps schools design:

​

• Multi-Year, Spiraled Curricula: Strategic, step-by-step academic frameworks that align AI literacy with student developmental milestones.

• Institutional Efficiency Frameworks: Practical systems that leverage AI to streamline administrative and grading workflows, ensuring that time saved is directly reinvested into student outcomes.

• Measurable Benchmarks: Clear, metrics-driven evaluations that protect academic integrity while embracing modern technological literacy.

Jim’s mission is simple: to help schools move past defensive "banning and policing" and transition to an offensive strategy of empowerment, preparing students to thrive in an AI-driven world without losing their humanity. It allows for more connection

Generative Thinking: A New Approach 
to AI in the Classroom©

By Jim Roche
Abstract
For many students (and skeptical educators), AI is often viewed as the end of inquiry—a shortcut to an answer or a finished essay. Generative Thinking flips the script on that idea.
A Generative Thinking Curriculum centers the student as the active agent. Instead of passively receiving information, students learn to use AI as a catalyst for deeper thought and exploration. This article explores how an AI's initial output isn't a final answer, but rather a springboard for accelerated learning, iterative problem-solving, and enhanced critical thinking. 
This highly practical paper equips educators with the tools to create a multi-year scaffold for Generative Thinking. It provides ready-to-use strategies, including sample lesson plans tailored for elementary, middle, and high school levels, demonstrating how AI can act as a collaborative brainstorming partner.
The promise of AI is instantaneous access to virtually unlimited knowledge and analysis. The danger of AI—certainly in an academic setting—is that students passively and uncritically accept the “wisdom” given to them. Students and teachers often view AI as an easy way out: an answer and essay-writing machine; the end of inquiry. By the time our current K–12 students are parents and mid-career adults, AI will have a profound impact on their lives. They will be flooded with information, data, and analysis on a scale that is unimaginable today. The most successful members of their generation will be the ones who can take the endless piles of data and synthesize meaningful theories, concepts, and businesses. Educators must ask: “How do we build schools that teach the skills and abilities needed to succeed in the 2050s, 2060s, and beyond?”
Paradigm shifts are nothing new to education. Once, teachers relied on repetition, drill, and memorization, but they now focus on critical thinking and experiential learning. AI necessitates another shift. Schools must embrace Generative Thinking. Just as calculators did not erase the need to teach basic arithmetic, and spell and grammar checkers did not eliminate the need to learn those topics, AI will not replace basic skills and critical thinking. However, success in the middle of the 21st century will require something more—a skill set best described as Generative Thinking.
Generative Thinking flips the script on AI, and the student becomes the generative agent: not just analyzing and consuming information, but producing new ideas, frameworks, or questions that have not been asked before. The algorithm becomes a "brainstorming buddy." AI’s first answer is not the final word; it is the starting point of a student-led collaboration with AI. Generative
Thinking converts AI into a virtual experiential learning tool. We need to teach students to take back the driver’s seat from AI, to have conversations, and to brainstorm with AI. Educators must begin teaching how to use AI creatively and actively.
Scaffolding Generative Thinking (Elementary and Middle School)
Generative Thinking is not an assignment or even a class. It is a new approach to problem-solving—a method to enhance critical thinking. Like any skill set, teaching Generative Thinking is best achieved by scaffolding.
AI should be initially introduced to grade-schoolers as a chat buddy—a place for interactive discussions. Teachers should focus on helping students ask good queries and then meaningful follow-up questions. The emphasis will be on students chatting with AI, not asking one-time questions. They might read a story or watch a short video about butterflies. Next, the teacher opens a chatbot on the smartboard, and the students can begin asking questions about butterflies. They are encouraged not to just “shout out” predetermined questions but to read and think about the previous answers and then ask for more information about those ideas.
Elementary school teachers should also conduct assignments where they discuss the difference between good human questions and good AI questions. "What’s your favorite book, TV show, or video game?" are good human questions, but not good AI questions. They should also stress that students should chat with AI and ask meaningful, follow-up questions.
Middle school students should be assigned to have AI produce a result (e.g., developing a name for a cultural phenomenon or historical era), then ask AI to improve and refine the answers over several iterations. Students are graded not just on the result, but on how well they formulate questions and how actively they steer the AI discussion. By the time students reach high school, they are asked to use AI as their brainstorming buddy.
For example, after studying Colonial America, students ask AI to create three choices for a descriptive name for the American Colonial Era from 1700–1750. The student then picks their two favorites and asks AI to combine aspects of those two. AI is then asked to produce better names. The student must then go through three more iterations of improving the name, highlighting other things they have learned about that era. Once the students have completed four iterations, they choose the best name and give a short oral presentation explaining why they chose it. The focus here is on not accepting the first response as “the answer,” but learning to push AI to go deeper and do better—using AI to improve the student’s decision-making and reasoning skills, rather than having AI replace those skills.
Creating a Generative Thinking Curriculum
Creating a Generative Thinking Curriculum for the upper school requires more than swapping out lesson plans. It too must be scaffolded. Before attempting advanced lessons in Generative Thinking, students need to successfully complete beginner and intermediate lessons. Scaffolding in social studies might include the basic, intermediate, and advanced assignments below.
Upper School Scaffolding
Intermediate Assignment: After learning about the Revolutionary War, students are asked to write an in-class outline for an essay on the war. In the next class, they have AI write an outline for the same topic. (Multiple unique queries and detailed questions would be required for a good grade.) Finally, students must integrate their outline with the AI’s. They also submit the prompts that they used in research.
Advanced Assignment: Students are told that they work for a nonprofit organization that has just received a multimillion-dollar grant to fight the homeless crisis in a nearby city. They are tasked with writing a short proposal to the head of the charity recommending that the grant be used to develop an individual program. It should target a particular challenge facing the homeless population or a specific community (single mothers, children, veterans, etc.). It may focus on any relevant issue, from improving/expanding shelters, nutrition, drug addiction, health care, or employment opportunities.
Students are given one class period to research their program using AI. They are instructed to define the scope of the problem they are solving (e.g., how many homeless children are in their city) and to provide statistics to show the impact of homelessness on the target population. Then they should investigate what is currently being done to address this situation and what the unique challenges are in providing these services to the homeless. Next, the students look at what has been done in other places to address this issue and discover which approaches have been the most effective. After their research period, for homework, they must condense their results into one page of notes (not outlines or essays, but notes). Then, in the next class, they will use computers without internet access to write a white paper proposing their program to help address the homeless crisis. The grade for the assignment will be based primarily on the depth of their research as well as the complexity and nuance of their program. Extra credit will be given for particularly creative solutions that combine aspects of previously successful programs or suggest a novel approach.
Conclusion
Most educators first encountered AI with the introduction of ChatGPT in the fall of 2022. It burst on the scene as schools were still struggling to overcome COVID-19. Understandably, the initial reaction was defensive: "How do we prevent cheating and fraud?" While those solutions are still evolving, we have moved on to the next question: “How do we use AI to teach more efficiently?” How can teachers employ AI in lesson planning, course development, and other areas? As we continue working on those questions, we must also shift the paradigm of the role of AI in education. Teachers need to create a thoughtfully scaffolded Generative Thinking Curriculum that opens new doors to creativity, critical thinking, problem-solving, and knowledge accumulation. AI is a tool to expand the horizons of our students, not to limit their abilities and initiative. Generative Thinking will teach them the skills to become the leading thinkers of their generation.

and gaining more time to do what they want, when they want. Ironically, AI well-used can give more time to be with family and friends and picks up the repetitive busy work most people

​

Jim is available for a 15-minute consultation by phone or text at 201-240-0033 or jroche@aiedconsultants

​