First Draft Thinking

First draft thinking is Heather Schwartz’s term in What do students lose when they rely on AI for homework? for the protected first attempt a student makes before using AI. It includes the blank-page work of organizing information, forming an initial argument, trying a problem, combining skills, and discovering what the student does not yet understand.

The concept does not reject AI As Tutor. Its claim is about order. AI can explain, refine, compare, or extend after a student has done enough initial cognitive work, but when it supplies the first solution or draft, it can turn learning into passive consumption and deepen AI Shortcut Risk.

First draft thinking therefore connects classroom design to Self-Directed Learning, Learning How To Learn, and Human Judgment Under AI. Teachers can preserve it through AI-free time, teacher-led instruction, and supervised independent practice; students preserve it by delaying AI long enough to experience productive struggle rather than only answer retrieval.

Teaching students to ‘be better than a robot’ adds a complementary writing-center version through Christy Gerdhary. Once students have human material to work from, they can use AI to remediate, transform, or collaborate transparently, then judge whether the result still carries their own memory, language, and purpose. This extends first draft thinking into AI Writing Pedagogy rather than replacing it.

Are humans losing the ability to think for themselves? adds Steve Shaw’s broader “think first” rule. Shaw allows students to use AI and uses it daily himself, but he tells students to generate their own ideas before prompting. That makes first draft thinking a defense against Cognitive Surrender as well as a classroom writing practice.

Key Claims

• The first attempt is not a disposable step; it is where students practice synthesis, recall, transfer, and error detection.
• AI assistance is safer after the student has produced a draft, hypothesis, partial solution, or explicit confusion.
• Explanation quality does not guarantee learning if the student skips the reasoning that the explanation was meant to support.
• AI-free classroom time can be a practical scaffold for first-draft thinking because AI is difficult to exclude from home internet use.
• The goal is productive friction, not helpless frustration: AI can still enter after the learner has done enough thinking to use the help actively.
• In writing classes, the protected first draft can become the material students compare against AI transformations, color-coded collaboration, and final artifacts that exceed generic output.
• The same sequence applies outside homework: thinking before prompting helps keep AI from becoming the user’s first settled judgment.

Connections

• Heather Schwartz and RAND - source speaker and study context.
• Steve Shaw, Cognitive Surrender, and Artificial Cognition - Wharton decision-making extension added by Marketplace Tech.
• AI As Tutor - constructive AI role when sequencing is right.
• AI Shortcut Risk - failure mode when AI replaces the first attempt.
• Self-Directed Learning and Learning How To Learn - student capacities trained by first-draft work.
• Learning Experience Design - classroom and product design can create space for productive struggle.
• AI Use Pacing and Human Judgment Under AI - deciding when to slow down or refuse AI output is part of the skill.
• Christy Gerdhary, AI Writing Pedagogy, and Transparent AI Use - writing-class extension added by the January 15 Marketplace Tech episode.