Personal Research Project
MeaningAware
AI Safety Research
A personal exploration into whether meaning in text can be measured externally and deterministically — independent of the AI models being evaluated.
This project serves as a technical case study and self-directed learning exercise in computational semantics, AI alignment, and software architecture. Built on the Universal Meaning Equation (UME), it investigates structural omission detection, bias measurement, and narrative drift analysis.
Research Question
Can an external, model-agnostic measurement layer detect structural omissions and bias in AI-generated text — without relying on another AI model?
Status: Early-stage research. Proof of concept. No commercial activity, no clients, no revenue, no company.