The clay we don't shape

I want to share something that I’ve been grappling with for a while now: the use of AI in planning and implementing Project-Based Learning. 

The truth is that I might regret writing this article. My perspectives and experiences with AI and technology are still evolving and I am skeptical of anyone who claims to know everything or have all of the answers in this space. But I can reflect on where I stand now and my observations as things unfold. 

This is what I’ve been thinking about lately. 

An Undeniable Reality. 

When AI started to hit it’s big boom a couple of years ago, my first instinct was to play with the ChatGPT models and see how it could be helpful for teachers to use when planning Project-Based Learning (PBL). Every school in the world struggles to find enough time for teachers to plan quality instruction and anything that could help teachers plan faster and more efficiently is clearly beneficial. 

After experimenting for a bit, finding some productive ways to prompt the AI and even play with a few groups of teachers during PBL planning meetings,, I found that ChatGPT gave some good ideas. It could come up with a passable driving question and integrate Danish learning goals for each subject and grade level. The teachers were surprised that a simple chat-bot could provide so many of the ideas that would take them weeks to think of and decide on. 

It’s now 2026 and things have changed rapidly. A few weeks ago I sat with a teaching team who at one point, was extremely skeptical about using AI in their planning. Now, they all use it for planning on nearly a daily basis. They’re not alone. One study done in 2025 found that 30% of teachers were using AI in planning lessons- saving them about 6 weeks of work per year. In 2026, that figure was up to 60%. 

From my perspective, this has happened very, very fast. I have never seen a new platform or technology be adopted so quickly by individual teachers or by entire schools. Maybe it’s because ChatGPT (and other models) are so readily available and even the free versions are decent in helping with lesson planning. Maybe it’s because it matches the need for exactly what teachers spend a lot of time on. 

I want to say upfront that I am not anti-AI. I use Claude nearly every day and I’m fascinated by what this development in technology will do.

That said… I believe we have a responsibility to also be skeptical. And to notice what is happening when we use a new technology. To hold our attention to the consequences even though it may not be popular and it may feel like resisting something new for only the sake of resistance. I’ve asked myself this question often: Am I skeptical of AI because I just don’t like change? 

I doubt it… but it’s possible. 

No matter how I feel about it, the reality is that AI is already making a big impact on the education community and it’s not going away any time soon. 

Planning PBL and AI Models.

It’s common for me to receive a message in my inbox from a new tech start-up or someone working with a platform that uses AI to help teachers plan Project-Based Learning. Typically they either want me to push their platform to the schools I work with or they want feedback on how to make their models better from a PBL perspective. I take these meetings because I love understanding what’s happening in the PBL-world and I am genuinely interested in tools that can help teachers do PBL better. 

The reality is that nearly all of these meetings leave me with a sinking feeling in my stomach. 

It’s not because they’re bad or because the people behind them are awful. Usually it’s the opposite. I have a deep respect for people who are leveraging technology to make the classroom different for kids. 

But each platform I’ve seen over the last 5 years, has been pretty much the same. There is a way for teachers to manage the different groups, products and learning outcomes in a project and an AI feature that can help plan entire PBL projects or give sparring on a specific aspect of PBL. A teacher could type in a question like: what kind of products could students make that matches the question? And like in the experiments with the early versions of ChatGPT, the suggestions were fine. 

Maybe that word described the feeling in my stomach: fine. 

Would students be making something? Yes. 

Was there a question driving the project? Yes. 

Would the students be showing their work to someone else? Yes. 

I try to imagine myself in the shoes of a teacher. I’m busy, have very little time to plan PBL and I need to get a project together to start next week. 

Would these platforms be helpful? Yes. 

They make planning more efficient, some of the ideas better than pressured teachers can come up with and make it possible for someone without any prior knowledge or experience with PBL develop a recipe to implement a project. 

It’s the Ikea version of lesson planning. 

Is an Ikea chair better than no chair at all? Of course. 

But is something lost when we settle for the Ikea version? Absolutely. 

Molding The Clay. 

I was a PBL teacher at High Tech High for six years and I can tell you every single project I did with my students during that time. 

Not because they were all great and perfect, but because in the planning process, my hands were in the clay. Scooping up the cold, wet blob. Pumping the spinning wheel with my foot. Pinching my fingers to raise the edges, smashing the clay down when it rose too high or when I was generally dissatisfied. I sat next to my teaching partner who also was molding. Sometimes on the same wheel as me, sometimes with our legs touching but working side-by-side on shaping matching things. 

The Perfect School Project started as a casual conversation in our office. Legs up on the desk, a hot coffee in hand and my teaching-partner talking about his frustration with choosing a school for his daughter to attend. I mean, is there such a thing as the perfect school? He said with exasperation. He saw a gimmer in my eye. That was our next project. Our students had been complaining about our school a lot lately. There weren’t so many sports or activities they could join as a “normal” school and would PBL prepare them for college and their future? The Perfect School Project became a way for us to explore that question together. We started on our own research to become familiar with different types of schools, what other teachers were doing and how we could share what the students determined with an audience that would be really interested in their thoughts. One colleague had a brilliant idea: our students could run a workshop at the Deeper Learning Conference and share their ideas with educators from all over the world. 

Project Avielle was born from a need to capture the hearts of a difficult class and work on something that mattered significantly to them and to our local community. As we sat planning the project, we could only think of one day when our entire class of 9th graders was interested in school: December 14th, 2012. The director of our school informed us that there had been a shooting in Newtown, Connecticut and that one of our colleagues had lost a close friend’s daughter in the shooting. Our entire class was silent. Then they started to ask questions. Why would this happen? Who would do this? We spent the rest of the day watching the news and trying to understand. The project would be about designing a park to honor Avielle Richman and we got to work as a teaching team to understand park-design. We talked with our colleague and tried to understand more about Avielle and what would capture our students best. We arranged a call with Avielle’s dad and he agreed to talk with our students. I cried after we closed the video call. The project was already more meaningful to me than I could put into words- and we hadn’t even started. 

In almost every single project we planned, there was a process. Perhaps one of the most important parts of that process was us as teachers, falling in love and investing in the project before it ever reached our students. Every conversation we had about the question, the product, the audience- gave us a deeper connection to why it was important to do. Although time consuming and at times really frustrating, it also gave us ownership. 

I’ve learned that ownership in PBL from a student’s perspective often takes the shape of: “Your project… our project… my project.” Our end goal is always that students feel the project is theirs. The work is theirs. The impact is theirs. 

That ownership starts with the teachers though- what they’re passionate about, what they have invested time and energy into. One of the biggest complaints that I’ve heard from Danish teachers working in large teams with PBL is the lack of ownership they feel when they plan projects together. They want to contribute with ideas and feel like the project is theirs. 

And students feel that. 

They feel the difference between a project that has been downloaded and a project that has been intentionally designed. 

Recently though, I observed students working on a science project that had been created by a national science organization and distributed to schools. It included driving questions for each sub-topic, a product to build and daily lesson plans. That students were largely ambivalent about the project. They did the tasks but had a difficult time seeing a purpose for the work beyond the fact they would participate in a competition to see who had the best ideas. The teachers loved the project. When I asked them what was best about it, they told me that because it came with all the plans, PBL was easier. They felt more confident and had more energy to actually be in the project. 

It is easier to buy a mug off Amazon. It is more efficient, faster and you can use it the day it comes. 

But there is something unexplainable about making a mug yourself. It requires something. An investment, grappling. Love. 

Values.

I honestly don’t know how to make sense of planning PBL projects with AI. 

I would hate to be looked at as someone who wants to prevent teachers from using AI when it is helpful and makes their job easier. I know we don’t live in a utopianic world where teachers have massive amounts of time to collaborate, reflect and design projects. I would also never advocate for a situation where teachers should not use AI because of some set of principles that no one can live up to. 

We’re in the real-world, Loni. 

I keep getting drawn back into the values in which I see teachers making decisions to use AI- and I think it could be applied to anyone who is using AI-models right now more generally. I’ve seen it said many times: What you spend time on, shows what you value.

When I see most teachers use AI to plan PBL, it’s in service of something that is quick, efficient and always available. It doesn’t depend on time, surplus, support-hours, money or another person to collaborate with. It also doesn’t require a deep understanding of PBL- the elements, practices or mindsets. 

Which means the values teachers are operating under are: speed, efficiency, availability and least amount of friction. 

The way I see it though, is that those are the exact things that Project-Based Learning is trying to do the opposite of. 

I don’t see schools choose PBL because it’s the easiest option or because it gets kids to learn faster and more efficiently. 

I see schools choose PBL because of a desire to have students go deeper. To create more opportunities for investment and motivation. I hear schools talk about the importance of building stronger communities that consist of more collaboration and working together. I observe schools get excited about working in ways that students can be creative, use their hands and contribute in ways that make a real difference in their local contexts. 

My worry is that using AI in PBL planning will not make us better at those things. 

The thing that will make us better at those things is creating the conditions for teachers to also experience them on a daily basis. 

I don’t think we become better at facilitating investment by skipping the investment ourselves. 

I don’t think we get more creative and skilled at building by delivering packed lessons without engaging in the process ourselves. 

I don’t think we improve at collaborating and helping others to collaborate by planning alone. 

I don’t think we go deeper with learning in the classroom without committing ourselves to the deep work of understanding. 

Of course, I could be wrong. 

What Now?

Using AI in PBL might be the revolution that makes PBL accessible and possible for even more schools. It might give students even more opportunities to engage in meaningful and deep learning. It might empower a generation of activists, critical thinkers and do-ers that the world so desperately needs right now. 

When I see AI provide this kind of transformative power in schools for both children and adults, I will have no problem reflecting on that and sharing my evolving thoughts. 

I have seen AI be helpful to teachers in PBL planning when: 

• Teachers already have an understanding of good PBL and can evaluate the quality of AI input 

• AI is used as a sparring partner with specific elements of PBL (help us twist the question to use with older students) 

• AI is prompted with ambitious PBL definitions that will push project design 

• It is used to produce resources (like a worksheet for a feedback protocol) that support the element of PBL but are time consuming for teachers to create 

• Teachers are using AI when planning together- not in insolation 

Even given the benefits, the pit in my stomach just won’t go away though. 

I wonder sometimes if I’m just romanticising how PBL should be and not dealing with the reality of how things are in schools? Or if I’m resisting technology because I feel old and long for the analog days of the 1990’s? Or if I am holding on to a much too idealistic version of PBL that is impossible for schools to live up to? Am I scared that AI will replace me as a PBL partner to schools? I will admit that all of these things could be true. 

As it stands now, I have three main worries: 

• The quality of PBL when planned with AI models (whole-project design and sparring on ideas) is low. The more we accept these outputs as “good” PBL, the more we water-down the potential and power of PBL. 

• Without molding the clay in project planning, how can teachers model the values of PBL? 

• If PBL becomes faster and more efficient with AI planning, will PBL become just a series of delivered projects and lose its impact on whole-school development? 

Here’s the thing though. 

I never started working with Project-Based Learning because it was easy. 

For me, PBL has always been a vehicle to transform schools in a deep and profound way. To shift the values, organization and practices to be a better place for all kids and to shape the world we live in. 

None of that is easy. Or efficient.

It happens through friction and investment and love.

And AI simply cannot do that. 

I am happy to hold the space of healthy skepticism for now.