Guided Scholar · Responsible AI

AI that supports learning, not replaces it.

Guided Scholar was built on a straightforward principle: a tool that does the thinking for the student has failed its purpose. Every design decision reflects that principle.

Our Position

What we believe about AI in the classroom.

AI has a legitimate and valuable role in education — but that role is not to produce student work. It is to make the feedback loop faster, the evaluation criteria more visible, and the revision process more actionable. When those things happen, students learn. When AI completes the task for them, they don't.

Guided Scholar was designed around this distinction from the ground up. The system provides structured feedback, rubric-aligned evaluation, and specific revision guidance. It does not write essays, complete assignments, or generate content on behalf of students. The student does the work. Guided Scholar helps them do it better.

This is not a policy we added after the fact. It is the foundational design decision that every feature in the system was built around.

Core Principles

How these principles shape the product.

Principle 01
Students do the work.

Guided Scholar never generates student content. Every word in a submitted draft was written by the student. The system responds to what the student produced — it does not produce it for them.

Principle 02
Teachers remain instructional leaders.

Guided Scholar gives teachers visibility into student progress and revision behavior. It does not replace teacher judgment — it supports it by surfacing information that would otherwise require reading every draft manually.

Principle 03
Feedback serves revision.

Feedback in Guided Scholar is always specific, actionable, and connected to what the student actually wrote. The goal is not a score — it is a student who understands what made their draft stronger and can apply that understanding to the next attempt.

Principle 04
Evaluation is transparent.

Every rubric category is explained in plain language. Students see not just what they scored but why — which criteria were met, which were underdeveloped, and what a stronger response would look like. Evaluation should build understanding, not just produce a number.

Specific Commitments

What Guided Scholar will and will not do.

Student data is not used to train AI models. Submissions, drafts, and revision history are used to generate feedback for that student. They are not used to improve or retrain any AI model. Rubric and scoring calibration is conducted using purpose-built benchmark essays, not student work.
The system does not generate student content. Guided Scholar provides feedback, revision suggestions, and structural guidance. It does not write paragraphs, complete sentences, or produce assignments on behalf of students.
Feedback includes one example rewrite per submission. Each feedback response includes a single example showing the student's original sentence alongside a stronger version. This is a teaching feature, not content generation. The example is scoped to one sentence and is intended to illustrate a revision principle the student can apply independently.
Revision visibility is real, not estimated. The teacher dashboard shows actual draft text, word counts, and structural changes between submissions, not inferred or approximated revision behavior.
ACT scoring is validated against human-graded essays. Guided Scholar's ACT rubric scoring was validated against six sets of essays graded by humans using official ACT criteria. 66 of 66 essays scored within one point of the human-graded score.
Data retention is defined and limited. Student submission data is retained for six months, archived for an additional six months, and then permanently deleted. Schools and districts receive clear documentation of this policy.

What Guided Scholar is not.

Not an open AI chatbotGuided Scholar is a structured educational workflow tool, not a general-purpose AI assistant. Students interact with specific assignment templates and receive rubric-aligned feedback — not open-ended AI conversation.
Not an essay generatorThe system cannot produce student work. This is not a policy limitation — it is a design limitation built intentionally into the product.
Not a replacement for teachersGuided Scholar surfaces information and provides feedback at scale. Teacher judgment, instructional leadership, and direct student relationships remain central to how the system is designed to be used.
Not unrestricted AI accessStudents interact with Guided Scholar through structured workflows tied to specific assignment types. The system is purpose-built for educational contexts, not general AI access.

Questions about how Guided Scholar handles AI in your school or district?

peter@guidedscholar.ai