Learning Assurance Framework
Final submissions no longer prove learning.
Project Gnoletheia documents the learning path instead of policing the final product, so transparent AI use is rewarded, honest students are protected, and degrees stay credible.
Truth made visible and anchored.
Problem
Assessment is built on a broken assumption.
That the document a student hands in mirrors what they learned. Generative AI broke that link, and the tools meant to repair it made things worse.
Polished work in minutes
Generative AI produces fluent homework and papers on demand. A finished document says nothing about who learned what.
Detectors cannot be trusted
Similarity checks and AI detectors are unreliable, biased against non-native writers, and trivially gamed. Guessing is not assurance.
Honest students pay the price
Blanket suspicion poisons the classroom. Institutions fight AI instead of teaching students to use it well.
Approach
Don’t detect cheating. Document the journey.
Gnoletheia records how work unfolds, draft by draft, prompt by prompt, and verifies the process rather than judging the product. Three layers work together, with humans deciding every case that matters.
Input layer
Students and teachers set up honest work from the start. Every step leaves an xAPI trail, so the path is visible by construction.
Agentic analytics
Private analysis reads the trail, flags meaningful anomalies, and builds a picture of real engagement over the semester.
Verification
AI triages, humans decide. Routine cases resolve automatically; anything ambiguous goes to a reviewer with full context.
Pilot
Starting where the problem lives.
The first deployment runs with the Faculty of Informatics Engineering at Lattakia University, in real courses, with real grading stakes.
Partner with usPricing
Universities pay. Students just learn.
The licence is bundled with registration each semester, so no student is priced out of fair assessment. AI inference is metered separately to keep the base price honest.
For universities
Annual licence in seat bands (500 / 2,000 / 5,000), platform plus base AI inference included.
- One-time onboarding: setup, staff training, LMS integration
- Monthly AI overage bundles priced to scale
For individuals
Base usage with selected AI models and tools, free forever.
- Pro plan ($5–10/month) for frontier models and premium tools
- Pay-as-you-go bundles when you exceed the base
Team
Built by someone who sat in those seats.
A Lattakia University graduate, now building the tool he wishes his own faculty had.
Software engineer, Lattakia University graduate. Co-founder of workers.vc and maintainer of stift.secundus.studio.
Contact
Help prove learning, not paperwork.
We are looking for pilot faculties, quality-assurance committees, and LMS teams. Project status: pre-pilot.