← Back

AI Disclosure

Effective: TBD

Peaks Education uses third-party AI models to generate explanations,

quizzes, flashcards, summaries, and study suggestions. This disclosure

explains what we send, who processes it, and what to expect.

Models we use

Provider Models Used for
Anthropic Claude family Tutoring chat, multi-step study agents
Google Gemini family Schedule extraction, syllabus parsing, grading

We may swap or upgrade models periodically. Functionality and the

categories of data sent stay the same.

What we send

For tutoring requests we send:

  • Your message or question.
  • Relevant context from materials you've added to your library

(syllabi, notes, prior questions).

  • A system prompt describing the tutoring task and academic-integrity

rules.

For schedule extraction we send:

  • The image, PDF, or text you uploaded.

For syllabus parsing we send:

  • The uploaded syllabus file content (text extracted from PDF/DOCX).

We do not send your password, payment details, or unrelated personal

information.

What providers do with that data

Both Anthropic and Google offer enterprise terms that exclude customer

content from training their general-purpose models. We use these

terms.

Both providers may retain content for a short window for abuse

prevention; that retention is governed by their own policies.

Limitations of AI output

  • AI output can be wrong, outdated, or made-up.
  • AI output may produce content that conflicts with your instructor

or your school's guidelines.

  • AI output is not a substitute for course instruction or for doing

your own work.

You are responsible for verifying anything you rely on for academic

work.

Your choices

  • You can avoid AI features entirely by not opening tutoring chat,

not running quizzes, and not using auto-extraction. The non-AI

features (manual notes, library, schedule view) remain available.

  • If your school requires an "AI off" mode for your account, contact

support@peakseducation.com.

Changes

We will post material changes here with a new effective date.

Version v1.0 — effective 5/2/2026.