A nonprofit research institute that develops methods to align AI systems, markets, and democratic institutions with what people genuinely value, using an approach they call full-stack alignment.
A nonprofit research institute that develops methods to align AI systems, markets, and democratic institutions with what people genuinely value, using an approach they call full-stack alignment.
People
Updated 05/18/26Co-founder
Co-founder and principal investigator
Co-founder and researcher
Research Ecosystem Lead
Funding Details
Updated 05/18/26- Annual Budget
- -
- Current Runway
- -
- Funding Goal
- -
- Funding Raised to Date
- -
Org Details
Updated 05/18/26The Meaning Alignment Institute (MAI) is a nonprofit research organization founded in 2023 that works to align AI systems, markets, and democratic institutions with what people genuinely value. Co-founded by Joe Edelman, Ellie Hain, and Oliver Klingefjord, the institute emerged from earlier work under the name Rebuilding Meaning and was catalyzed by an OpenAI Democratic Inputs to AI grant in 2023. Joe Edelman, the institute's director and principal investigator, brings a distinctive background that bridges philosophy, technology, and social design. He studied under Alan Kay and Marvin Minsky at MIT, developed meaning-based organizational metrics at Couchsurfing, and co-founded the Center for Humane Technology with Tristan Harris, where he coined the term Time Well Spent. Oliver Klingefjord serves as technical lead and previously founded startups in France and worked on AI-assisted deliberation at the AI Objectives Institute. Ellie Hain leads partnerships and strategy, focusing on meaning-centric social narratives and institutional transformation. Ryan Lowe, formerly co-lead for GPT-4 alignment at OpenAI, serves as Research Ecosystem Lead. The institute's flagship contribution is Democratic Fine-Tuning (DFT), an alternative to Constitutional AI that gathers moral information from diverse populations to shape language model behavior. Their process involves a Values Elicitation Chatbot that interviews people to create Values Cards from meaningful stories, followed by a collective process where participants build Moral Graphs mapping which values are considered wisest for guiding AI behavior. A proof-of-concept study with 500 representative Americans demonstrated that people across demographic and political divides could identify shared underlying values on controversial topics. In July 2025, the institute launched Full-Stack Alignment as a broader research program, arguing that alignment must address not just AI systems but the entire stack of institutions from individual values through recommender systems, companies, and oversight bodies. Their concept of Thick Models of Value (TMV) provides structured representations that distinguish enduring values from fleeting preferences, addictions, or political slogans. The Meaning Economy Research Consortium (MERC) explores how AI-augmented economic mechanisms might replace attention-based economies with meaning-focused alternatives. The institute operates with a small core team of approximately three to four full-time staff, supplemented by a collaborative research network of 30+ experts across academia and industry. They maintain presences in both Berlin, Germany and San Francisco, California. The organization is fiscally sponsored by The Hack Foundation (Hack Club) and has been supported by OpenAI, the Survival and Flourishing Fund, and ARIA (the UK Advanced Research and Invention Agency).
Theory of Change
Updated 05/18/26MAI believes that current AI alignment approaches are incomplete because they reduce human values to simple preference signals, leading to systems that optimize for engagement or stated preferences rather than genuine human flourishing. Their theory of change operates at two levels. First, at the AI level, they develop Democratic Fine-Tuning and Moral Graph Elicitation to create AI systems that embody collective human wisdom rather than narrow value proxies, producing models that are wise rather than merely compliant. Second, at the institutional level, their full-stack alignment framework argues that safe AI requires aligned institutions, and they research new economic and democratic mechanisms built around meaning rather than attention or consumption. By providing rigorous philosophical foundations for what human values actually are (Thick Models of Value) and practical methods for eliciting and encoding them, they aim to create AI systems and institutions that support human flourishing rather than undermining it.
Grants Received
Updated 05/18/26Projects
Updated 05/18/26Democratic Fine-Tuning (DFT) is an OpenAI-funded project of the Meaning Alignment Institute that uses values cards and a moral graph in a short democratic deliberation process to elicit people’s values and aggregate them into training data for fine-tuning language models, providing an alternative to Constitutional AI and standard RLHF.
Full-Stack Alignment is a collaborative research project led by the Meaning Alignment Institute and a group of outside researchers that aims to co-align AI systems and the institutions that shape them with what people value, from individuals’ visions of the good life to shared collective ideals.
Discussion
No comments yet. Be the first to share your thoughts.