Stages of AI Adoption from AI skeptic to AI champion part 3 of 4

This article is part of the series “From AI Skeptic to AI Champion,” which explores how educators can move from fear and uncertainty toward wise, ethical, and human-centered AI adoption.

Teachers do not adopt new technologies all at once. Some teachers are early adopters. They are curious, experimental, and willing to try something new before all the answers are clear. Others are more cautious. They wait, observe, ask questions, and look for evidence. Still others are skeptical, not because they are lazy or resistant, but because they care deeply about students and have seen enough failed innovations to know that not every new thing is a good thing. This is especially true with Generative AI.

AI is not simply another educational tool. It raises deeper questions about learning, authorship, truth, assessment, student formation, teacher identity, privacy, equity, and wisdom. When teachers hesitate, we should not immediately interpret hesitation as resistance. Often, hesitation is a sign that teachers understand the moral weight of their work. At the same time, schools cannot remain stuck in hesitation.

Students are already using AI. The workplace is already changing. AI tools are already being embedded into search engines, writing platforms, productivity suites, learning systems, and creative tools. Educators need a roadmap for moving forward wisely. That is why I have found it helpful to think in terms of AI Adoption Stages.

The purpose of the AI Adoption Stages Model is to provide educators, school leaders, and policymakers with a clear framework for understanding and supporting teachers’ evolving relationships with AI. The model is based on interviews with hundreds of K-12 teachers conducted in 2024-25 and identifies five stages of adoption (with a pre-adoption Stage 0: AI Skeptic), each marked by specific mindsets, behaviors, questions, and challenges. The goal is not to shame teachers into moving faster. The goal is to help each teacher responsibly, conscientious, and courageously take the next wise step.

Stage 0: AI Skeptic – Reframe

This is the Pre-Stage. The AI Skeptic has little to no experience with GenAI and may have significant concerns about its use in education. Some skeptics are strongly opposed to AI. Others are not opposed, but they are unconvinced. Their posture is often, “Not yet.” Sometimes it comes across as, “Never.”

Stage 0 is not necessarily a problem. Skepticism can be healthy. In fact, schools need thoughtful skeptics. They ask questions enthusiastic adopters may overlook. They notice risks. They slow down careless implementation. They protect students from being treated as test subjects for every new technology. But skepticism becomes dangerous when it hardens into bias, fear, or permanent inaction.

The invitation at Stage 0 is to reframe AI from threat to tool. That does not mean ignoring the risks. It means becoming willing to investigate whether AI can be used for good in education. At this stage, teachers should read, watch, listen, ask questions, and observe someone they trust using AI. They might register for ChatGPT or another approved tool and try one simple task. Not a classroom redesign. Not a student-facing activity. Just one small experiment. The key question is: Can you show me that AI is worthy of my trust, time, and effort?

Stage 1: AI Novice – Understand

The AI Novice has begun to explore. At this stage, teachers may use AI for simple personal tasks: planning a vacation, generating dinner ideas, asking for a summary, brainstorming a home project, or using AI as an advanced search companion. They may not yet understand prompt engineering, model limitations, hallucinations, bias, or effective educational use. But they are experimenting. This matters because experience changes attitude.

Many fears about AI are abstract until teachers actually try it. Once they do, they often discover two things at the same time. First, AI is more powerful than they expected. Second, AI is less perfect than it appears. Both discoveries are important. The goal at Stage 1 is to understand. Teachers begin learning how AI responds to different prompts. They discover that vague prompts produce vague answers and that better prompts produce better responses. They learn to give AI a role, goal, context, audience, and format. Teachers at Stage 1: AI Novice “use it at home but not at school.” The key question is: What can I learn about AI by using it?

Stage 2: AI Practitioner – Personalize

The AI Practitioner begins applying GenAI directly to the work of instruction. This is where many teachers first experience meaningful professional benefit. AI can help create lesson plans, discussion questions, rubrics, differentiated readings, quiz questions, parent emails, letters of recommendation, examples, non-examples, and project ideas. At this stage, AI is usually working for the teacher, not directly with students. That makes Stage 2 a relatively safe and productive stage for many schools. Teachers can experiment in a sandbox. They can test outputs, revise them, and decide what is usable.

For example, a teacher might ask AI to create three versions of a reading passage at different reading levels. Or a high school teacher might ask AI to generate five discussion questions for a novel, then revise the questions to better match the class. Or a professor might ask AI to draft a rubric and then improve it based on course outcomes. Another teacher may use AI to create images or videos for a lesson.

At Stage 2, teachers often discover an initial return on investment. AI can save time, reduce repetitive work, and help teachers move from a blank page to a usable first draft. But this stage also requires professional judgment. AI output must be checked. Teachers must verify accuracy, revise tone, remove bias, align with standards, adapt to students, and make sure the work reflects their own instructional purpose. AI can draft. AI can suggest. AI can organize. But the teacher remains responsible.

Teachers are also learning that outputs depend directly on inputs. For example, instead of asking, “give me a lesson plan”, a teacher might ask: “Act as a fifth-grade teacher. Create a 30-minute introductory lesson on equivalent fractions for students who struggle with multiplication. Include one visual analogy, two guided practice problems, one common misconception, and one exit ticket.” This kind of guided experimentation builds confidence. The key question is: How can AI help me become a better teacher?

Stage 3: AI Manager – Delegate

The AI Manager begins delegating selected teaching functions to AI. At this stage, AI may help with evaluation, scoring, analysis of student work, or communication. Unlike Stage 2, where AI helps the teacher prepare materials before instruction, Stage 3 uses AI to help the teacher respond to student work after instruction. AI may interact with student work, but the teacher still mediates the interaction. AI becomes a teaching assistant, not the teacher.

For example, a teacher can run student assignments through AI to generate possible constructive feedback. The teacher then reviews the AI-generated response and decides what to use, revise, reject, or rewrite. Some teachers may use AI feedback as a starting point and heavily personalize it. Others may reject the feedback entirely and write their own. Some may be tempted to copy and paste the AI feedback verbatim without any review. That temptation is the real risk of Stage 3.

An AI Manager could also use AI to evaluate and score student work with or without a rubric. Imagine an advanced Scantron machine. Traditional Scantrons can grade multiple-choice exams without teacher judgment because they merely detect whether the correct bubbles have been filled in. AI can fulfill this basic function, but it can do much more. With teacher guidance, AI has the potential to evaluate more complex student products such as essays, graphs, problem-solving explanations, lab reports, discussion posts, concept maps, slide decks, business plans, portfolios, and other artifacts that reveal student thinking.

Multimodal AI expands this capability even further. In theory, multimodal AI can evaluate creative student work such as digital or physical art, music compositions, architectural designs, videos, storyboards, and poetry. But as one might predict, the more complex, original, and creative the product, the less formulaic the evaluation can be. A multiple-choice answer can be marked right or wrong. A math problem can be checked for accuracy. But a poem, a song, a painting, or a design often requires human interpretation, context, and cultural understanding.

At Stage 3, teachers are learning how to delegate. But in doing so, they are also testing the boundaries of trust and responsibility. On the one hand, teachers worry that AI-generated feedback will be inaccurate, inappropriate, or inferior to human review. On the other hand, teachers may believe that evaluation and feedback are core irreplaceable tasks of the educator that should not be delegated to AI at all. Both concerns are worth taking seriously.

Evaluation is not merely a technical act. Feedback is not merely information transfer. When teachers respond to student work, they are doing more than identifying errors. They are communicating what matters, what growth looks like, what the student has understood, and what the student should attempt next. A good teacher knows when a student needs correction, when a student needs encouragement, when a student needs a challenge, and when a student needs grace. AI may assist with this work, but it cannot fully understand the student, the relationship, the classroom context, or the larger purpose of the assignment. This is why mediation is the defining feature of Stage 3. The AI Manager does not blindly accept AI output. The AI Manager supervises it. The teacher asks: Is this accurate? Is this fair? Is this aligned with the rubric? Is this developmentally appropriate? Does this sound like me? Will this help the student grow? What does the AI miss because it does not know the student? How can I connect with my students? How can I express grace when grading student work? How can evaluation and feedback function as a way to know and see my students?

Used wisely, AI can help teachers notice patterns, identify strengths and weaknesses, compare student work to criteria, and respond more efficiently to student learning. Used carelessly, it can produce generic comments, questionable scores, inaccurate judgments, or impersonal responses that weaken the teacher-student relationship. The key question is: How can AI delegation elevate teacher empathy and sustain student growth?

Stage 4: AI Partner – Design

The AI Partner designs educational environments in which AI becomes part of the learning experience itself. This is different from simply using AI to prepare materials or generate feedback. At Stage 4, students may interact with AI in role-plays, simulations, debates, design challenges, coaching cycles, and inquiry-based learning experiences. AI becomes a partner in the learning environment, while the teacher remains the instructional leader.

For example, students studying ethics might role-play a difficult moral dilemma with AI taking on one perspective. Students studying history might interview an AI simulation representing a historical figure, then critique the accuracy and limitations of the simulation. Students practicing leadership might rehearse a difficult conversation with AI, then reflect on their tone, clarity, empathy, and decision-making. Statistics students may get help on challenging materials from an AI tutor that is encouraging, engaging, and understands student motivation.

Research suggests that AI Partnership (Stage 4) in practice produces some of the strongest returns on investment because it approximates the benefits of individualized instruction, mastery learning, and immediate formative feedback (Kestin et al., 2025; Deng et al., 2025; Létourneau et al., 2025). These studies also reveal that students report higher engagement, increased motivation, in addition to significant achievement gains. Each of these cases also emphasize that the greatest instructional benefits come with intentionally designed, pedagogically robust, teacher-supervised AI partnerships. AI never replaces the teacher. It changes the role of the teacher.

The difference between Stage 3 and Stage 4 is subtle but important. In Stage 3, AI often provides support in a more unidirectional way: feedback, explanation, suggestions, or evaluation. And always through the mediation of the teacher. The AI does not interact directly with students. In Stage 4, the interaction becomes more dialogical, collaborative, immersive, and most importantly, direct and unmediated. They are thinking with, against, through, and beyond AI.

Consequently, this requires mature instructional design. The teacher must supervise the task, frame the learning, set ethical boundaries, anticipate problems, and require reflection. AI partnership does not mean teacher absence. It requires even stronger teacher presence. The key question is: How does a teacher’s role shift when partnering with AI in the classroom?

Stage 5: AI Champion – Transform

The AI Champion helps others move forward. At this stage, teachers are not only using AI in their own practice. They are mentoring colleagues, modeling responsible use, leading professional development, building communities of practice, contributing to policy conversations, conducting research, presenting, writing, coaching, and helping schools develop shared wisdom. AI Champions are important because adoption cannot depend only on outside experts. Schools need internal leaders who understand the local culture, students, teachers, mission, and constraints.

But the best AI Champions are not reckless enthusiasts. They remain humble. They continue learning. They tell the truth about AI’s strengths and weaknesses. They are patient with skeptics because they remember that they too were once beginners. The key question is: How do I invite others on this journey of discovery, curiosity, and courage?

Every Stage Matters

One mistake school leaders can make is using adoption stages as a ranking system. That is not the purpose. Teachers may occupy more than one stage at the same time. A teacher may be an AI Practitioner when it comes to subjective inquiry and an AI Partner when it comes to dialectic discussion topics. Another teacher may use AI for lesson planning but not yet for assessment. Another may be ready to design student-facing AI activities in one subject but not another.

Nonetheless, higher stages typically require greater investment and greater risk. Schools and districts also differ in their risk tolerance. Some ban AI entirely. Some allow teachers to use it but prohibit student use. Some allow student use within defined boundaries. These institutional contexts shape what teachers can reasonably do.

The goal is not to force everyone immediately to Stage 5. The goal is to help every educator take the next faithful step. For the AI Skeptic, the next step may be watching a trusted colleague use AI. For the AI Novice, it may be learning how to write a better prompt. For the AI Practitioner, it may be using AI to create differentiated materials. For the AI Manager, it may be using AI to co-evaluate student work. For the AI Partner, it may be building a simulation or role-play that students play with to deepen learning. For the AI Champion, it may be mentoring others with patience and humility.

Generative AI is changing education. But teachers do not need to be dragged into the future. They can be invited, supported, challenged, equipped, and encouraged. The journey from AI Skeptic to AI Champion is not a race. It is a developmental journey that starts with a single step.

References

Deng, R., Jiang, M., Yu, X., Lu, Y., & Liu, S. (2025). Does ChatGPT enhance student learning? A systematic review and meta-analysis of experimental studies. Computers & Education, 227, Article 105224. https://doi.org/10.1016/j.compedu.2024.105224

Kestin, G., Miller, K., Klales, A., Milbourne, T., Ponti, G., & Woolf, B. P. (2025). AI tutoring outperforms in-class active learning: An RCT introducing a novel research-based design in an authentic educational setting. Scientific Reports, 15, Article 17458. https://doi.org/10.1038/s41598-025-97652-6

Kinnen, M. (2024). AI learning framework [Unpublished framework].

Létourneau, A., Deslandes Martineau, M., Charland, P., Karran, A. J., Boasen, J., & Léger, P.-M. (2025). A systematic review of AI-driven intelligent tutoring systems in K–12 education. npj Science of Learning, 10, Article 29. https://doi.org/10.1038/s41539-025-00320-7

Microsoft & LinkedIn. (2024). 2024 Work Trend Index annual report: AI at work is here. Now comes the hard part. Microsoft WorkLab. https://www.microsoft.com/en-us/worklab/work-trend-index/ai-at-work-is-here-now-comes-the-hard-part

World Economic Forum. (2025). The Future of Jobs Report 2025. World Economic Forum. https://www.weforum.org/publications/the-future-of-jobs-report-2025/