Differentiated Instruction With AI in 9-12 Classrooms: What to Use and What to Skip

Seventeen sophomores are staring at a three-page excerpt from The Great Gatsby, and eight of them have completely checked out by the second paragraph. My co-teacher and I looked at each other across the room, knowing the exact scale of our failure. We had handed the same dense text to our English honors kids, our on-grade-level students, and our English language learners who were still wrestling with conditional verbs. Our old routine of scrambling at 11:00 PM to rewrite passages manually was dead, but adapting materials for AI differentiated instruction for 9-12 classes required a complete operational overhaul.

Three Reading Levels from One Source Text

The fastest way to fix mixed-ability reading blocks is to feed a vetted, copyright-compliant text into an engine like Diffit and request three distinct lexile bands at once. The free tier offers a limited number of generations per month, while paid accounts run around $15 per user per month. It shines when you need printable and digital versions for mixed-ability classes with minimal setup, though the leveling remains an approximate guide rather than a formal psychometric assessment. When leaning on this for high school text work, you have to watch out for the generated comprehension checks, which are often the weakest link.

Prompt: Act as a high school English teacher. Read the attached excerpt on industrialization and generate three distinct reading versions: a supported version at a 6th-grade lexile with glossed vocabulary, a core version at a 9th-grade lexile, and an extension version at an 11th-grade lexile with analytical questions.

Output received: Three clean versions with vocabulary lists, plus five multiple-choice questions per tier. The supported text was genuinely accessible, but the core version included a distractor question that flatly contradicted a primary source fact.

Teacher edits:

1. Deleted the factually incorrect multiple-choice question from the core tier.
2. Replaced two definitions in the glossed vocabulary list that relied on overly abstract terms.

If you prefer an all-in-one suite approach for text work, Eduaide.Ai provides a clean, uncluttered interface at roughly $10 per user per month with a daily generation cap. It handles worksheet generation efficiently, though its template set is narrower than heavy hitters like MagicSchool. For deep structural document work, I often pivot to Claude, which costs around $20 per user per month on its paid tier and handles long-context inputs remarkably well. Claude avoids the overly generic prose typical of smaller tools, but it lacks classroom-specific templates and must be kept strictly to teacher-facing tasks to protect privacy.

Inline Feedback Without the Copy-Paste Loop

Moving between a source text, a prompt window, and a student drafting document eats up precious planning periods. Brisk Teaching solves this by operating as a browser extension that brings AI assistance directly into Google Docs, Slides, and web articles. It costs approximately $10 per user per month with a tight free tier. The primary strength is eliminating the endless copy-paste cycle, giving you fast inline feedback on student drafts with a gentle learning curve. Because it is a Chrome-first workflow, your district IT department will need to vet the extension permissions and data scope.

When using Brisk for differentiating writing prompts in a senior composition class, the secret is targeting specific rubric criteria rather than asking for general praise. Never paste identifiable student work into unvetted consumer tiers; always check your district data privacy policies. If you need a more robust, district-approved environment for student-facing AI interaction, SchoolAI provides teacher-controlled student spaces designed around robust safety defaults. It runs around $15 per user per month and offers full teacher visibility into live student sessions, though it requires broader school buy-in and relies heavily on quote-based pricing.

Building Low-Stakes Engagement and Formative Checks

High schoolers can smell a time-wasting worksheet from across the hallway. Curipod helps turn standard lecture topics into interactive, student-response lessons in minutes, costing about $12 per user per month with a restrictive free tier. It works exceptionally well for quick checks for understanding and high student engagement right out of the box without requiring student accounts for basic use. The catch is that it focuses purely on lesson delivery rather than deep offline planning, and it completely falls apart if your classroom device cart is uncharged.

When designing a five-minute exit ticket set through these platforms, specificity is your shield against generic AI sludge. You must explicitly prompt the generator for targeted cognitive tiers so you do not end up with three variations of the exact same recall question.

Prompt: Create a three-question exit ticket for a 12th-grade government lesson on executive orders. Include one strict recall item, one application scenario involving a modern environmental regulation, and one reasoning item regarding constitutional limits.

Output received: Three clean questions formatted nicely. However, the generator snuck in a second recall question disguised as an application task.

Teacher edits:

1. Deleted the duplicate recall question that the AI added on its own accord.
2. Rewrote the application scenario to feature a local zoning dispute rather than a federal abstraction to match our current unit anchor.

Navigating IEP Drafts and Question Banks Safely

Differentiated instruction in high school also means handling the administrative mountains of IEP present-levels statements and accommodation tracking. When drafting these documents, you must use aggregated, anonymized data only. Ask the AI for observable, measurable statements, but replace every single generated number with your own hard progress monitoring data before a case manager ever sets eyes on the document.

Similarly, when turning a complex PDF article into a differentiated question bank, feed in non-identifiable source material. Ask the tool to tag items by cognitive difficulty and require it to state the specific misconception behind each distractor. Always spot-check the answer key by hand, as numerical calculations and data-interpretation items generated by LLMs are notoriously unreliable. If you are building out your entire tech stack for the term, review our broader guides, including AI accommodations for middle school classrooms to see how scaffolding evolves across grade bands, and check our master index of the best AI tools for teachers in 2026.

FAQ

Can AI do differentiated instruction at 9-12 level?

It can produce a first draft and a structure. The level judgement, the standards check and the tone are teacher decisions.

Which tool is best for differentiated instruction?

Diffit on the evidence in this article, with Brisk Teaching as the second opinion. Neither replaces the review step.

What is the biggest risk?

A confident, plausible output that nobody checked. The five checks below exist for that reason.

Sources

• Diffit Product Overview and Pricing: https://diffit.me/
• Brisk Teaching Chrome Extension Features: https://briskteaching.com/
• SchoolAI Safe Workspace Documentation: https://school.ai/
• Curipod Interactive Lesson Platform: https://curipod.com/
• Eduaide.Ai Resource Library Specifications: https://eduaide.ai/
• Anthropic Claude Business and Education Guidelines: https://anthropic.com/claude