As a district administrator and professor of educational leadership, Dr. Andy Szeto has a unique, ground-level view of the challenges and opportunities facing schools today. He has spent his career teaching and advising the next generation of school leaders while navigating the practical realities of instructional improvement. In his writing, he cuts through the noise surrounding educational technology to focus on a persistent, often-maligned cornerstone of teacher growth: professional development. Moving beyond the familiar satire of chaotic workshops, Dr. Szeto explores how artificial intelligence can address the deep, systemic failures in how we support our educators. This conversation delves into the practical applications of AI that can reduce administrative burdens, provide sustained and personalized coaching, and create a more meaningful, data-informed culture of professional growth, all while navigating the critical ethical considerations.
Your article opens with the all-too-familiar “Linda the literacy expert” satire. Beyond the humor, what specific, systemic PD failures does this highlight, and how would an AI-powered system begin to address the chaos and disconnect that teachers experience in such sessions?
That “Linda” character resonates so deeply because she embodies the fundamental disconnect in so much of professional development. It’s not just about a disorganized presenter; it’s about a system that often treats PD as a compliance checklist rather than a genuine opportunity for growth. The core failures are a lack of context, a one-size-fits-all delivery, and a complete absence of sustained support. “Linda” doesn’t know the district, the teachers, or the students, yet she arrives with a pre-packaged, generic solution. An AI-powered system directly counters this by starting with the user. Instead of a single, generic workshop, an AI platform can help a leader design a session built on teacher self-assessments or even anonymized student performance data. It can generate differentiated materials on the fly, so while one group of teachers might be working on foundational strategies, another could be tackling advanced applications of the same concept. This immediately dismantles the “elbow partners” and “dance for literacy” chaos by replacing it with relevance and a clear connection to the teachers’ actual classroom realities.
You mention a “quiet mountain of invisible work” in PD planning. For a school leader feeling overwhelmed, what are the first one or two concrete steps they could take with a tool like Gamma or Canva, and what kind of time savings could they realistically expect?
That “mountain” is so real, and it’s where good intentions for PD often go to die. For a principal or an instructional coach who feels like they’re drowning in logistics, the first step is to automate the foundational tasks. Don’t start by trying to build a complex, district-wide system. Start with your next PLC meeting or staff workshop. Instead of staring at a blank slide deck, give a tool like Gamma a simple prompt: “Design a 60-minute professional development session for high school history teachers on implementing inquiry-based learning, including learning objectives, an agenda, and three interactive activities.” Within a minute, you have a professional-looking, standards-aligned draft. The time savings are immense. What might have taken three or four hours of painstaking work—finding templates, writing copy, aligning to standards—is reduced to about ten minutes of refining and personalizing. This isn’t about replacing the leader’s instructional vision; it’s about clearing away the clerical underbrush so they can spend their precious time focusing on the most important part: designing a truly engaging and human-centered learning experience.
The concept of Novobo, an AI “mentee” that teachers train together, is fascinating. Could you walk us through how a session like this might work in practice, and what specific tacit skills you’ve seen teachers uncover when they are forced to explicitly teach an AI?
Imagine a group of veteran math teachers in a room, tasked with teaching Novobo, this AI agent on a screen, how to explain the concept of slope. One teacher might start by saying, “It’s rise over run,” and uses her hands to make a climbing motion. The AI tries to mimic it but gets it wrong. This forces a conversation. Another teacher jumps in, “No, we have to be more precise. Let’s use the graph. We need to show the change in y first, then the change in x.” As they work together, using their voices and gestures to guide the AI, something magical happens. They begin to articulate the hundreds of micro-decisions and non-verbal cues they use in the classroom without even thinking about them. One teacher might realize she always uses a specific tone of voice when she gets to the critical part of an explanation, a tacit skill she’d never consciously considered. By having to make their internal expertise external and explicit for the AI, they not only strengthen their own understanding but also build a shared language and practice as a team. It transforms PD from passive reception to active, collaborative construction of knowledge.
You liken AI to a “GPS for professional growth.” Beyond just recommending courses, how does an AI-powered system actively close the implementation gap for a teacher’s long-term goals? What does that ongoing support look like weeks or months after the initial workshop is over?
The GPS analogy is key because a GPS doesn’t just give you a map and wish you luck; it provides turn-by-turn directions and reroutes you when you hit a dead end. That’s precisely where traditional PD fails and AI can succeed. Let’s say a teacher attends a great workshop on student-centered discussions and sets a goal to increase academic talk in her classroom. An AI-powered coaching platform can close the implementation gap by providing that ongoing, gentle guidance. Two weeks later, she could upload a short video of her class discussion. The AI coach might provide asynchronous feedback, saying, “You did a wonderful job of facilitating the initial conversation. Next time, try posing a follow-up question that asks students to build on a classmate’s idea. Here are three potential question stems.” This is not evaluative; it’s supportive. Months later, the system might notice she’s consistently using this skill and suggest a new, related goal or connect her with another teacher in the district who is an expert in Socratic seminars. It’s this sustained, personalized, and data-informed feedback loop that transforms a one-day workshop into a long-term journey of professional growth.
Many schools struggle to evaluate PD beyond satisfaction surveys. Using Guskey’s five levels as a framework, how can AI automate data collection to measure actual changes in practice and student outcomes? Please share an example of what that might look like for a PLC.
The reliance on “smile sheets” is one of the biggest weaknesses in PD evaluation. Guskey’s framework gives us a much richer way to think about impact, and AI can be the engine that collects the necessary data. Let’s take a PLC of elementary teachers focused on a new phonics intervention. At Guskey’s Level 1 (Participant’s Reactions), an AI can instantly generate and analyze a feedback survey. But for Level 4 (Use of New Knowledge and Skills), things get interesting. Teachers could upload their weekly lesson plans, and an AI tool could be trained to scan them for evidence of the new phonics routines. It could provide immediate, non-judgmental feedback to the PLC on the fidelity of implementation. Then, for Level 5 (Student Learning Outcomes), the system could analyze the results from a common formative assessment on decoding skills, correlating the data with the implementation evidence from the lesson plans. The PLC would get a report showing, “The classrooms where the intervention was used at least four times a week saw an average 15% greater growth in decoding accuracy.” This gives them actionable evidence to refine their practice, moving evaluation from a feel-good exercise to a powerful tool for instructional improvement.
You rightly point out risks like the digital divide and data privacy. What specific policies or conversations must a district leadership team have in place before piloting an AI coaching tool to ensure teacher agency is protected and the technology promotes equity rather than widening gaps?
This is the most important conversation, and it has to happen first. Before any district signs a contract for an AI tool, the leadership team needs to co-author a transparent and robust data governance policy with their teachers’ union and a committee of educators. This policy must explicitly state that all data collected through AI coaching platforms is for formative, professional growth purposes only and will not be used for formal evaluation. It needs to be crystal clear about who owns the data, where it is stored, and how it is protected. Teacher agency is paramount; educators must have control over their own data and feel that the tool is a supportive mentor, not a surveillance camera. The second, equally critical conversation must be about equity. The leadership team must conduct an equity audit before a pilot, asking hard questions: Do all our schools have the necessary bandwidth and devices? What is our plan for providing ongoing training and support, especially for teachers who may be less comfortable with technology? If we don’t address these issues proactively, AI tools will inevitably become another resource that widens the gap between our most and least-resourced schools. True progress requires building a foundation of trust and equitable access from day one.
What is your forecast for the future of AI in professional development over the next five to ten years?
My forecast is one of optimistic pragmatism. I believe we will see a profound shift where AI becomes the indispensable “co-pilot” for instructional leaders and coaches. It will handle the 80% of logistical and data-analysis work that currently consumes so much of their time—scheduling, generating first drafts of materials, analyzing student work samples, and tracking progress toward goals. This will free up the humans in the system to focus on the 20% of the work that truly matters: building relationships, providing nuanced and empathetic feedback, fostering creativity, and leading complex, collaborative inquiry. The future isn’t about replacing coaches with algorithms. It’s about empowering our best educators with intelligent tools so they can do more of the deeply human work that inspires real, lasting growth in our teachers and, ultimately, our students.
