Three weeks ago, two public schools in Houston launched a bold academic experiment.
The schools will split the school day this year, with mornings devoted to intensive traditional academics and the afternoons centered on projects that will teach “human-centric skills” like critical thinking, collaboration, and adaptability.
The idea is to ensure students master both the academic basics and the higher-order skills necessary to stand out in a world that is increasingly reliant on AI. On its website, the school district said the new program “expands the focus to include critical thinking, problem-solving, real-world experiences, and the ability to navigate an increasingly complex, technology-driven world.”
At first blush, this seems like an interesting approach to a real problem. I love that a major U.S. school system is trying to adapt to a new and ever-changing reality with foresight and thoughtfulness. Rather than banning AI or pretending generative AI shouldn’t change how schools operate, Houston is facing the cold reality: with AI predicted to consume most routine work tasks, being competitive in the workforce will mean mastering higher-order reasoning skills, offering creative problem solving, and strong emotional intelligence.
Houston’s willingness to pivot and adapt is commendable, but as any veteran teacher will tell you, it’s a big leap from a press release to an effective implementation of a new way of doing things.
U.S. schools may learn a lot from the effort in Houston. However, there are some aspects of the effort that I’ll be watching closely to see how they’re handled.
Rather than banning AI or pretending generative AI shouldn’t change how schools operate, Houston is facing the cold reality: with AI predicted to consume most routine work tasks, being competitive in the workforce will mean mastering higher-order reasoning skills, offering creative problem solving, and strong emotional intelligence.
You Cannot Think Critically On Nothing
The basic idea behind Houston’s switch is to separate academics – which will be taught “intensively” in the morning hours – and the learning of “human-centric skills” – which will be taught in afternoon classes, seminars, field trips and group projects. As an idea, it sounds fine, but it may be fatally flawed.
As cognitive scientist Daniel Willingham has shown, critical thinking and problem solving are not independent skills that can be taught (or learned) in isolation.
In his book, Why Don’t Students Like School?, Willingham argues that the process of thinking is tightly bound to domain knowledge. In other words, you can’t engage in deep critical thinking about an issue if you don’t know much about that issue. An afternoon spent learning about “moral reasoning” won’t achieve much if the students lack the historical, scientific, or cultural knowledge needed to puzzle out a moral dilemma in the real world. For skills like this (and especially for critical thinking) context matters. Facts matter. Prior knowledge is required.
If Houston schools treat the morning’s academic blocks as separate from the afternoon skills building, then this model may not fare well. Skills like analytical reasoning and problem solving flow from deep content knowledge. If these two things are disconnected in the classroom, students could end up practicing high-level thinking on low-level knowledge.
An afternoon spent learning about “moral reasoning” won’t achieve much if the students lack the historical, scientific, or cultural knowledge needed to puzzle out a moral dilemma in the real world.
The Risk Of ”Watered Down” Curriculum
Beyond the need for domain knowledge, there is another pitfall worth looking out for: higher-order skills like “collaboration” and “adaptability” are all too easy to water down into low-rigor activities. We’ve all probably taken part in a group project where a few people do most of the work, but everyone gets the credit. That’s a simple example, but it’s worth thinking about as long as afternoon workshops or project time could end up a waste without strong guardrails.
What do strong guardrails look like? As part of the afternoon sessions, teachers could require students to detail and explain their decision making, planning, approach and revisions, for example. This would ensure all students are actively participating and prevent this new approach from becoming fluffy busywork.
Beyond the need for domain knowledge, there is another pitfall worth looking out for: higher-order skills like “collaboration” and “adaptability” are all too easy to water down into low-rigor activities.
Target ”Thinking About Thinking” First
Nevertheless, I think Houston’s approach is a good one. Of all the higher order skills needed to compete in an AI-fueled world, metacognition ranks among the most important. Houston teachers at these test schools should make sure that metacognition is baked into every afternoon session.
Simply, metacognition is thinking about our thinking. More deeply, metacognition is our ability to contemplate, assess and regulate our cognitive processes.
In a time when AI can almost instantly output homework answers, solve math problems, or even pen entire essays, students can skip over the heavy mental lifting of learning. This is called “cognitive offloading” and it’s a real problem.
But if taught metacognition skills, students will be able to recognize when AI is doing the thinking for them as opposed to helping them think more deeply about a topic. Furthermore, the skills they’ll learn to challenge their own thinking can be applied to AI outputs. For example, a common question critical thinkers often ask themselves is, “Where did my logic break down?” Students trained to think this way could instead apply that question to AI – Where did the AI’s logic break down? – to better assess the validity and salience of AI-produced material.
Teaching metacognition will require students to explain their reasoning, defend their ideas, and verify AI material against real-world evidence. These skills could transform students from passive recipients of information into active auditors of knowledge.
Teaching metacognition will require students to explain their reasoning, defend their ideas, and verify AI material against real-world evidence. These skills could transform students from passive recipients of information into active auditors of knowledge.
Then Target Agency To Empower Students
This shift — from passive learners to active auditors — creates new opportunities for student agency. All too often, American public schools are set up to deliver instruction to students, rather than empowering students to seek out and manage their own learning. When an AI can do a student’s homework flawlessly and instantaneously, the old mechanisms of assessing student learning break down.
Houston’s plan has a solution to this problem baked in: Students will learn the “human-centric” skills while working on big projects. According to the district, one of those projects is “TBD The Musical.” This project would have students working together as a creative team to create an entire musical from scratch “starting with no script, no songs, and no rules.” The district says “students learn to write scenes, compose original music, and choreograph dances through constant trial and error.”
This approach could be a vehicle to promote student agency as it shifts the student experience from receiving information to devising and executing a plan with others. This is the kind of learning project that many parents want their children to take part in, and it could be one of those memorable school projects that students carry with them for years.
However, there’s a risk. How will students write music if they’ve not learned the basics of melody and rhythm? How can they write a script if they’ve not been exposed to the fundamentals of storytelling? This gets back to what Willingham meant when he said higher-order thinking requires knowledge about the subject matter.
All too often, American public schools are set up to deliver instruction to students, rather than empowering students to seek out and manage their own learning. When an AI can do a student’s homework flawlessly and instantaneously, the old mechanisms of assessing student learning break down.
The Bottom Line For Parents And Educators
Houston’s recognition that school needs to adapt to the realities of AI is commendable. It acknowledges that the old model of schooling was designed for a time when information was scarce and learning facts was the most important objective. But today we live in a world where information is abundant, free and accessible 24/7. Teaching students how to navigate and succeed in this environment is the biggest question facing education today.
Is Houston’s plan the answer? From where I sit, the key to its success will be to what degree the morning academic sessions provide the content needed to executive the afternoon projects, and whether those afternoon sessions are held to a standard of cognitive rigor.
Already, press reports show parents are pushing back and questioning the new approach. I hope Houston leaders stick with it, revise as they learn, justify their decisions, and show everyone watching that they’re using the same high-order skills that they want to instill in their students.
This column first appeared in Forbes.
