ĀML™ — ĀRU Meaning Language™ is an open, meaning-native accountability layer for AI-generated interfaces that makes declared purpose, policy, authority, and evidence inspectable before and after rendering.
Modern AI can generate interfaces dynamically, but conventional web stacks do not inherently explain why a machine chose to show a particular element, which policy allowed it, or what evidence remains afterward. ĀML is a working research prototype exploring a new layer between machine intent and human-facing UI: an AI Interface Firewall™ that can evaluate meaning and policy before rendering and produce independently inspectable receipts and provenance.
ĀML is designed to complement—not automatically replace—HTML, React, Next.js, or existing frontend systems.
HTML tells the browser what to display. ĀML tells the system why it deserves to be displayed.
The project asks a broader technical question:
If AI is increasingly deciding what humans see, should those interface decisions carry inspectable meaning, policy, authority, and evidence?
ĀML's answer is yes—and the repository contains a working prototype of that idea.
- Live proof: https://aruintelligence.github.io/aml-core/proof.html
- Playground: https://aruintelligence.github.io/aml-core/playground.html
- View Meaning™: https://aruintelligence.github.io/aml-core/view-meaning.html
- Decision gallery: https://aruintelligence.github.io/aml-core/gallery.html
- Repository: https://github.com/aruintelligence/aml-core
- Publications: https://github.com/aruintelligence/aml-core/blob/main/PUBLICATIONS.md
The public implementation includes parsing and semantic structures, deterministic policy decisions, accessibility/consent/privacy-oriented controls, execution receipts, cryptographic verification, provenance, audit structures, protocol and schema work, RFCs, conformance fixtures, browser bridges, GitHub Actions integration, enterprise pilot material, and independent-verification tooling.
ĀML should be evaluated on reproducible technical behavior, not hype. The project does not claim that:
- ĀML is already a ratified Internet or web standard;
- its current attention/restoration inputs are objective measurements of cognition or wellbeing;
- a prototype policy guarantees ethical outcomes;
- cryptographic integrity proves that an underlying human-impact judgment is correct;
- existing browsers or web frameworks must be discarded.
Those boundaries are intentional. The project is trying to make a new class of interface accountability technically testable.
For technology media:
Can AI-generated interfaces explain why they exist? ĀML proposes a new accountability layer.
For developer audiences:
A policy firewall for AI-generated UI: ĀML adds inspectable meaning before rendering.
For AI governance audiences:
From model output to human interface: ĀML makes UI decisions auditable and verifiable.
For security audiences:
ĀML explores cryptographic receipts, provenance, and policy enforcement for machine-generated interfaces.
For standards audiences:
ĀML publishes an open protocol, RFC, schema, and conformance path for meaning-native interfaces.
“The web gave developers View Source. ĀML explores View Meaning™.”
“AI can propose an interface. ĀML can make its meaning, policy, authority, and evidence inspectable.”
“The goal is not to hide machine decisions behind a prettier interface. The goal is to make the decision itself inspectable.”
ĀML overlaps with accountable AI, generative UI, agentic interfaces, AI governance, explainability, policy enforcement, provenance, consent, privacy, accessibility, interface security, semantic analysis, cryptographic receipts, human-centered AI, and interoperability.
ĀML™ / ĀRU Meaning Language™ is developed by ĀRU Intelligence Inc.
Commercial, enterprise, OEM, licensing, strategic, and media inquiries: Office@aruintelligence.com
Official branding and marks are governed separately from the repository's open-source code license. See TRADEMARKS.md, COMMERCIAL.md, and the official authorization documentation in this repository.