Projects

Our fellows develop higher-tier capacities through real projects. The projects serve the mission. The mission is not the projects. Each project below is named for the human capacities it develops — the tiers that AI cannot supply — not just for what it builds.

Tiers refer to the Irreducibly Human framework. Learn more at irreducibly.xyz

Agents

Dayhoff

Open-source AI agents for biological research, bioinformatics, computational biology, and public health.

Develops:
Tier 5 — Causal ReasoningTier 6 — Collaborative Synthesis

Madison

Open-source agents for marketing and branding, providing a framework for data-driven advertising intelligence.

Develops:
Tier 4 — JudgmentTier 5 — Causal Reasoning

Mycroft

Educational experiment using specialized AI agents for finance to analyze the investment landscape.

Develops:
Tier 4 — JudgmentTier 5 — Causal Reasoning

Popper

AI agents for computational skepticism, DevOps, and AI ops, including assertion verification that scans claims and flags suspect ones for expert review.

Develops:
Tier 4 — Judgment

Research

Medhavy

Adaptive AI-powered learning platform with multi-armed bandit pedagogy engine.

Develops:
Tier 4 — JudgmentTier 6 — Collaborative Synthesis

AI for Good

Homes of Hope

Homes of Hope +1 is an eight-week AI+1 literacy and job-readiness bootcamp for girls in Hyderabad, India, developed by Humanitarians AI.

Lyrical Literacy

AI-generated songs and videos that activate the embodied and social mechanisms of language acquisition.

Develops:
Tier 2 — EmbodiedTier 3 — Social & Ethical

LoonNet

A Humanitarians AI for Good initiative built in partnership with a Loon Conservatory, whose first playable educational prototype is A Loon’s Lake, a four-chapter game.

Jobs

Become a Fellow

Fellows work on these projects while developing the judgment, causal reasoning, and collaborative synthesis that AI cannot replicate. The portfolio is evidence. The capacities are the product.

Humanitarians AI Lyrical Literacy Project