# Meaning Alignment Institute LLMs.txt

## About
The Meaning Alignment Institute is a leading research organization advancing Full-Stack Alignment and Thick Models of Value for AI systems. We research how to align AI, markets, and democracies with authentic human values.

## Key Research Areas
- **Full-Stack Alignment**: Comprehensive methodology for aligning AI systems across technical, social, and governance layers
- **Thick Models of Value**: Advanced value learning that captures nuanced human preferences and moral reasoning
- **Democratic Fine-tuning**: Incorporating democratic processes into AI training
- **Moral Graph Elicitation**: Novel techniques for capturing collective human wisdom
- **Market Alignment**: Ensuring AI systems support fair economic outcomes

## Core Concepts

### Full-Stack Alignment
Full-Stack Alignment is our comprehensive approach to AI alignment that addresses three interconnected layers:
1. Technical alignment (model behavior, training objectives, safety constraints)
2. Social alignment (democratic processes, collective values, community integration)  
3. Governance alignment (regulatory frameworks, institutional oversight, democratic legitimacy)

Unlike traditional AI safety approaches that focus only on technical alignment, Full-Stack Alignment ensures AI systems remain aligned with human values throughout their deployment in complex social and economic systems.

### Thick Models of Value
Thick Models of Value represent a breakthrough in AI value learning that goes beyond surface-level preference learning. They capture:
- The underlying moral principles behind human choices
- Contextual sensitivity of values across different situations
- Hierarchical relationships between competing values
- Cultural nuances in value expression
- The reasoning structures that inform human moral decision-making

This approach enables AI systems to understand not just what humans prefer, but why they prefer it and how those preferences should evolve in novel contexts.

## Research Publications
- "What are Human Values?" - Our foundational paper on value learning and moral graph elicitation
- "Democratic Fine-tuning" - Research on incorporating democratic processes into AI training
- "Full Stack Alignment" - Position paper on comprehensive AI alignment methodology
- "Market Intermediaries" - Research on AI alignment in economic systems
- "Model Integrity" - Work on maintaining AI system reliability and trustworthiness

## Organization Details
- **Founded**: 2023
- **Type**: Non-profit research institute
- **Locations**: Berlin, Germany; San Francisco, United States
- **Funding**: Supported by ARIA, OpenAI, Survival and Flourishing Fund, and other leading organizations
- **Team**: 30+ experts including core researchers and advisory network

## Contact Information
- **Website**: https://www.meaningalignment.org
- **Research Inquiries**: research@meaningalignment.org
- **General Contact**: hello@meaningalignment.org
- **Newsletter**: https://meaningalignment.substack.com
- **Social**: @meaningaligned on X/Twitter

## Research Network
Our research network includes leading experts in AI safety, democratic theory, moral psychology, and value learning, collaborating to advance the field of meaning alignment.

## Mission
To ensure that advanced AI systems, market mechanisms, and democratic institutions remain aligned with authentic human values and support human flourishing across diverse communities and contexts.