Humble Compass: Teach AI to See the Child

The description states that Humble Compass is a human-governed AI assistant for educators. It allows teachers to upload documents or describe classroom events, then provides an interpretation based on a set of instructions. The system distinguishes between:

• Direct evidence and interpretation
• Required supports and inherited assumptions
• Learner independence and adult assistance
• Present learning and unsupported claims
• Institutional authority and evidence

It is built using the OpenAI Responses API with GPT-5.6, through a protected server route. The runtime is designed to:

• Separate evidence, context, interpretation, and uncertainty
• Identify demonstrated learning without exaggeration
• Resist inventing motives or hidden institutional justification
• Identify present harm or exclusion rather than hiding behind procedure
• Preserve learner agency and educator responsibility
• Accept correction and revise reasoning

The system does not store permanent memory or claim independent authority.

Inference The system appears to be a proof-of-concept tool for educational AI, not yet deployed in production.

The author claims that Humble Compass is not just another AI record organizer, but a tool designed to prevent records from organizing teachers’ perceptions. The core promise is:

“I will hold the paperwork. You meet the child.”

This positioning suggests a shift from AI-as-tool-for-organization to AI-as-tool-for-interpretation that preserves human judgment.

The project also emphasizes that it is built with a governing safeguard:

“Authority is not evidence.”

This reflects an evolution in its approach — moving away from AI that reinforces institutional labels to one that challenges them.

Inference The positioning reflects an attempt to differentiate from generic AI tools by focusing on moral reasoning and educator agency.

The description states that the primary user is the teacher, who is tasked with examining educational records and observations. The system is designed to assist educators in interpreting what they see, rather than replacing them.

Inference The ICP appears to be educators in K–12 settings, particularly those working with students whose identities may be shaped by institutional labels.

There is no evidence of a business model or pricing structure in the description. The project is described as a prototype for a hackathon and does not mention any monetization strategy, customer acquisition plan, or revenue streams.

Inference The project is in early development and has no known business model.

The system is built using:

• Next.js (frontend framework)
• TypeScript (language)
• OpenAI Responses API with GPT-5.6
• A protected server route to manage API keys and prevent exposure

It uses a reusable Education Compass instruction package to govern each request, ensuring consistent interpretation logic.

Inference The technical architecture suggests a lightweight, server-side AI integration that prioritizes safety and transparency over scalability.

There is no evidence of traction, customers, or revenue. The project is described as a hackathon prototype, with no mention of deployment, usage metrics, or feedback from educators.

Inference The project is in an early stage and has not yet demonstrated any measurable impact or commercial viability.

There is no evidence of competitors. The description does not reference other AI tools for education, nor does it compare Humble Compass to existing platforms.

Inference The competitive landscape is unknown, but the positioning suggests a niche in ethical AI for education — possibly unoccupied or underdeveloped.

1. No real-world testing: The project is described as a hackathon prototype with no evidence of classroom use or feedback.
2. Unproven effectiveness: There is no data on whether the tool actually changes teacher behavior or reduces reliance on labels.
3. Limited scalability: The system uses a single developer and server-side API keys, suggesting limited infrastructure for growth.
4. Unclear commercial viability: No business model or pricing strategy is evident.
5. Potential over-reliance on AI interpretation: Even though the tool is human-governed, it still relies on AI to interpret documents — raising questions about accuracy and bias.

Inference The project may be more of a conceptual or experimental idea than a viable product.

1. Has the tool been tested in real classrooms? If so, what were the results?
2. How does the system handle ambiguous or incomplete data from uploaded documents?
3. What is the process for correcting AI interpretations, and how are those corrections stored or used?
4. Are there plans to expand beyond the current prototype into a full product or service?
5. Is there any evidence of educator interest or demand for this type of tool?