AI for Humanity
We believe the most valuable resource in the AI age isn't compute or data, it's our Benevolent Bandwidth. It's the human capacity to care, build, and ensure technology serves the many instead of the few.
We solve problems with transparent and open technologies that deliver public benefit by prioritizing safety, fairness, and trust.
We are building a world where the benefits of AI are shared by everyone, strengthening fairness, opportunity, and human agency.
We identify important problems and ideas that deliver public benefit which can be solved with technology
We identify, motivate and onboard technical advisors and contributors
We connect ideas with contributors
We execute on the ideas and connect users with the tools
Stewards of the foundation
Owners of projects and repos
Execution leaders
Associate Professor of AI in Climate Change, Georgetown University
Previously: VP of Data Science, Capital One
Founder, NdP coaching
Previously: VP of Community Investing and Impact, Capital One. COO, DataKind
Distinguished Professor of Marketing and Assoc Dean of Research, Northeastern University
Chief Marketing and Communications Officer, Prudential Financial
Previously: Head of Core Mrketing, CapOne
Managing VP of Data and Data Science, Capital One
Previously: Dir of Research, Travelers
Associate Dean and Program Director for Economics and Analytics, Boston College
Founder, The Benevolent Bandwidth Foundation
Head of Global Credit Infrastructure, Data and AI, PayPal
Head of Real Estate Solutions, Proxet. Managing Principal, Line Advisors. Attorney at Law
Previously: Co-Founder and CEO, LeasePilot
Director of AI, Capital One
Previously: Fermilab, Stanford Linear Accelerator Center
VP of Data Science, ANZ Bank
Previously: Fidelity Investments. Director of AI
AI Product Founder, HappyFin
Previously: Velocity Black and Capital One
Applied Scientist, Amazon AGI
Previously: Lead Machine Learning Scientist, Applied Research (LLMs), AI Foundations Team, Capital One
Senior Director of AI, Tredence
Previously: Sr Lead of AI,Fidelity Investments
Data Engineer and Analyst, Capital One
Previously: Data Analyst, Fragport AG
Member of Technical Staff, Nile
Previously: Data Science Manager, SiriusAI
Manager, Data Analytics & Engineering, Tesla
Previously: Data Scientist, eCare Services
BI Developer, Qvinci Software
Previously: Business Analyst, IBM. Senior Consultant, Systems Limited
Professor of Applied Analytics, Boston College
Director of AI, Capital One
Previously: Director of AI, Fidelity Investments
Vice President, Data Science, Fidelity Investments
Previously: Sr. Research Manager, IMRB International
Our projects represent a vision for using AI and technology to solve real-world problems. Each idea has been carefully vetted to ensure alignment with principles of public benefit, privacy protection, and open-source development. Active projects and our backlog span several key problem domains where technology can create meaningful impact. Our goal is to develop deep expertise in each domain over time, building specialized solutions that address systemic challenges through AI-powered tools.
Our initial focus areas span the following domains:
Over time, we aim to partner with non-profits, NGOs, and other public benefit organizations to jointly solve problems at scale using technology and AI, amplifying the collective impact on the communities we serve.
Problem domain: Access to Information
Too many important papers get lost in the noise. Most researchers and practitioners cannot reliably scan what is new recently in their area, find truly promising work, and trust that they did not miss something big. This agent helps with this problem by finding, ranking, and explaining recent AI papers on arxiv.org.
Problem domain: Environmental Impact
EcoTag is a smartphone app that empowers consumers to make sustainable choices by revealing the hidden environmental cost of their clothing. Using Optical Character Recognition (OCR) to scan garment tags, the tool calculates estimated lifecycle carbon emissions based on materials and origin. This provides shoppers with immediate, actionable data on the carbon footprint of their potential purchases.
Problem domain: Logistics
The Delivery Optimizer helps local businesses like florists and dry cleaners eliminate the waste of manual route planning. Owners simply upload a list of addresses to the web tool, which uses open-source algorithms to generate the most efficient delivery order. This streamlined process saves time and fuel while reducing the carbon footprint of local logistics.
Problem domain: Environmental Impact
People don't know if fixing old gadgets is worth the cost, so e-waste accumulates. This economic triage tool queries live marketplaces for part and resale prices to advise consumers whether to repair or recycle their old or broken electronics, helping reduce environmental waste while saving money.
Problem domain: Nutrition
Healthy marketing is deceptive and identifying ultra-processed foods on the NOVA scale is hard for consumers. This red light/green light scanner helps grocery shoppers avoid ultra-processed foods by using OCR and AI to read ingredients and classify foods, providing instant guidance on food quality.
Problem domain: Healthcare
Screening mammography has limitations, including false negatives and false positives. Individuals often lack a meaningful "second signal" on their results, or they experience anxiety from indeterminate findings. Second Look is a consumer-facing, AI-assisted interpretation companion that analyzes user-uploaded mammography images to identify regions that warrant attention and assign a coarse concern tier. It provides informed, conservative decision support to prompt professional follow-up, not to replace a doctor. The tool utilizes on-device inference for privacy and is available as a web or mobile application.
Problem domain: Access to Information
Beacon is a project designed to protect the elderly from malicious sites, emails, and links, including AI-generated content. This tool helps vulnerable users navigate the digital landscape safely by identifying and warning about potentially harmful content, reducing the risk of scams and phishing attacks that disproportionately target older adults.
Problem domain: Healthcare
Medical bills are riddled with errors and insurance denials are confusing, making it exhausting for patients to fight back. The Medical Bill Defender is a local-first, privacy-centric agent that helps patients in the US analyze their medical bills and Explanation of Benefits (EOB) documents. Users upload bill or EOB images or PDFs, and the tool checks CPT codes, flags errors, and automatically drafts appeal letters. The agent empowers patients to challenge incorrect charges and denied claims without the complexity that typically prevents them from taking action.
Join our community of contributors building AI tools for public good. Whether you're a developer, researcher, or enthusiast, there's a place for you.