Stephen Fitzpatrick and the AI Design Crisis Facing Schools
Source: Tom’s Takes: AI in Edu News, Views & Moves
Author: Tom Daccord
Original source: https://tomdaccordai.substack.com/p/stephen-fitzpatrick-and-the-ai-design
Published: 2026-06-11
Source type: essay
Private backup: the full article text is archived in the private repository at archives/articles/tomdaccordai-substack-com-stephen-fitzpatrick-and-the-ai-design.source.md. It is not published on the public Quartz site.
Summary
Tom Daccord uses a conversation with Stephen Fitzpatrick to argue that schools are treating AI too much as a control problem and not enough as a design problem. Policies, bans, detection tools, and permission categories may be necessary, but they do not answer the deeper instructional question: what thinking is the assignment supposed to build, and what evidence will show that students actually did that thinking?
The article emphasizes that students already use AI in many ways beyond full essay generation, including organizing notes, generating practice questions, studying for tests, revising writing, navigating math problems, and producing explanations. Because AI is also becoming embedded in search, documents, slides, learning platforms, and school management systems, Daccord argues that schools need principles that travel across tools rather than rules aimed only at named chatbots. The practical path he draws from Fitzpatrick is backward design: clarify learning goals, define where AI supports or undermines those goals, make student thinking visible before the final product, and keep human teachers close enough to interpret judgment, struggle, and growth.
Pull quotes
Design problem
“In an age of agentic AI, schools must redesign learning for a world where AI can manage the entire student workflow, not just provide answers.”
Assignment purpose
“The question is no longer simply whether students used AI. It is what kind of thinking the assignment was meant to develop, what evidence of that thinking remains visible, and where AI support crosses the line into cognitive offloading.”
Policy gap
“A gap Stephen identifies is that school AI policies often try to solve a design challenge with compliance language.”
Visible thinking
“AI did not break school assignments. It revealed what many assignments were failing to measure.”
Durability note
The article is durable because it connects several recurring wiki themes: AI-era assessment, tool-embedded AI, student offloading, and school AI implementation as instructional redesign. Its value is not a new taxonomy, but a clear synthesis of why compliance-only AI policy cannot substitute for redesigning tasks around learning purpose, visible process, and teacher judgment.
Big ideas
- District AI work is a long-term redesign project
- Learning still needs some struggle, even when AI can make things easier
Claims
- AI-assisted homework requires redesign, not just policing
- In an AI world, assessment should focus on watching students think
- District AI implementation needs living guidance and teacher-led redesign
- New AI capabilities will often reach schools through existing tools
Key evidence and examples
- Daccord frames the school AI problem as a shift from control to instructional design, because agentic AI can manage workflows rather than only produce isolated answers.
- Fitzpatrick describes students using AI as a tutor, study partner, organizer, explainer, quiz generator, research assistant, and feedback source, with uneven judgment about boundaries.
- The article argues that tool-specific policies age quickly when AI becomes embedded in Google Search, Docs, Slides, YouTube, Classroom-adjacent workflows, databases, and school management platforms.
- Daccord and Fitzpatrick treat final products as increasingly weak evidence of learning when AI can produce polished artifacts.
- The article recommends auditing major assignments for AI workflow vulnerability and deciding which parts of the learning process need to become visible.
- It argues that professional development should focus less on tool tours and more on backward design, cognitive friction, visible thinking, and evidence of student growth.
Education relevance
Highly relevant for K–12 AI policy, assessment redesign, academic integrity, teacher professional learning, and classroom assignment design. The article reinforces the wiki’s emerging position that schools need task-level redesign and visible evidence of thinking, not only rules about whether particular AI tools are allowed.