# We Must Stop Using AI to 'Level-Down' Our Students > [!question] Lingering Questions > - The piece never says when leveled text IS legitimately appropriate (decoding-stage readers, newcomer English learners) — where does the real professional line sit? > - It criticizes Diffit's core leveling function while praising the same tool for scaffolds. Does that soften a critique that should land on the vendor's headline feature? > - What's the professional-learning equivalent of leveled text — are AI meeting summaries doing the same thing to adults? > [!tip] Article Ideas > - "Chunk, don't change" as a documentation design principle — map the leveling-vs-scaffolding failure mode onto developer docs and onboarding content. > - Procurement rubric line item: "what human cognitive work does this tool absorb, and is that work the point?" > [!link] Cross Reference > - [[3. Resources/Permanent Notes/AI Should Raise Student Challenge, Not Lower Expectations]] — same domain, complementary claim > - [[3. Resources/Permanent Notes/AI Quality Gaps Are Policy Problems Not Capability Problems]] — the "blame training, not the tool" framing is worth arguing against using this existing note > - WCAG 3.1.5 recommends *supplementing* hard content, not replacing it — accessibility and this article agree, but "simplify for accessibility" is widely misread as license to replace. > [!quote] Vocabulary > - *Text leveling* — assigning students only texts at their measured "instructional level," rather than grade-level or aspirational texts. > - *Scaffolding* — temporary supports (chunking, vocabulary sets, think-aloud modeling) that keep the original text intact while making it accessible. ## Metadata - Source: [[3. Resources/Readwise/Articles/We Must Stop Using AI to ‘Level-Down’ Our Students|We Must Stop Using AI to 'Level-Down' Our Students]] - Publisher: EdSurge, part of a series on educator experiences supported by the Learning Commons; CC BY-NC-ND 4.0. ## Author's thesis Generative AI makes text-leveling — a discredited but historically labor-limited practice — nearly free to apply at scale, right as Science of Reading policy is legislating away from it. The tool isn't the problem; pointing AI at the text itself (simplifying it) removes the student's cognitive work and the learning with it, while pointing AI at scaffolds around an unchanged text (chunking, vocabulary, think-aloud scripts) preserves both. ## Key arguments (≤3) 1. **Removing the struggle removes the learning.** A dyslexic, ADHD eighth-grader given a third-grade summary of "All Quiet on the Western Front" showed no interest, waited to be told answers, and retained nothing — reversing to the real text with scaffolds produced classroom-level engagement three years running. 2. **The shortcut is seductive because it targets the real cost center.** Differentiation for students who don't get it after the third teaching is the most time-expensive part of the job; leveling tools promise to absorb exactly that cost, and did — along with the students' thinking. 3. **Leveling is an equity failure, not a neutral accommodation.** Giving some students the literary canon and others a machine summary of it converts a differentiation decision into an access decision. ## Claim atoms ### Removing effortful thinking removes the learning > [!quote] ID:rw_h1 > When we keep kids from actually doing work that requires effortful thinking, they don't learn anything at all. Where/why: opening, right after the district's AI leveling tool purchase. The mechanism the whole essay rests on — generalizes past classrooms to any tool that absorbs cognitive work along with the growth it would have produced. ### Struggling readers need a route, not an easier text > [!quote] ID:rw_h3 > My student didn't need a different text; she just needed a different roadmap and timeframe to get there. Where/why: "The Fallout," after reverting to scaffolded instruction on the original novel. Reframes differentiation from content substitution to path and pacing — the transferable design principle. ### Chunk the passage; never change the passage > [!quote] ID:rw_h4 > Use AI to chunk, not change, the passages. Where/why: "The Redemption," recommendation two of five sanctioned AI uses. The single most portable rule in the piece — six words that separate scaffolding from substitution, and apply unchanged to docs, onboarding, and training content. ### Leveled text creates inequity of canonical access > [!quote] ID:rw_h5 > why would we give some students access to the great literary canon, and others a robot-created summary? Where/why: "The Redemption," opening the responsible-AI-use section. Converts a pedagogy debate into an access-equity claim; the rhetorical-question form is quotable but unfalsifiable as stated — pair with data when reusing. ## Permanent candidates - **Automation that absorbs the effort absorbs the learning** — supported by H1; the general principle, scoped to cases where difficulty IS the point. → promoted. - **Scaffolds change the route; substitutes change the artifact** — supported by H4; the operational rule that follows from the general principle. → promoted. ## Processing checklist - [ ] Read Timothy Shanahan's "Leveled Reading: Leveled Lives" (cited as the author's reframing source) - [ ] Audit the last five AI-assisted docs or lesson artifacts I've produced for substitution vs. scaffolding - [x] Connect into the relevant MOC — [[3. Resources/MOCs/AI & the Future of Knowledge Work MOC]] ## Links - [[3. Resources/MOCs/AI & the Future of Knowledge Work MOC]] — the effort/substitution distinction generalizes directly to knowledge-work automation covered in this MOC