Bibliothèque

Towards Best Practices for Open Datasets for LLM Training

Building on community insights from 30 AI dataset experts, this research paper distills best practices for creating open datasets for LLM training. The paper is a collaboration between Mozilla and EleutherAI.

From Skin to Screen Bodily Intgrity in the Digital Age

Research examines the collection of “body-centric data” which has experienced a dramatic surge since the COVID-19 pandemic and the rise of sophisticated AI tools.

Incomplete Chronicles: Unveiling Data Bias in Maternal Health

An ethnographic research into the datasets powering maternal healthcare app “DawaMom” used across Zambia and other Southern African countries

Public AI

A vision for a robust ecosystem of initiatives that promote public goods, public orientation, and public use throughout every step of AI development and deployment.

Unveiling the Potential of a Conversational Agent in Developer Support: Insights from Mozilla’s PDF.js Project

This paper presents an investigation into whether AI can be leveraged to assist developers and guide open source community members. It introduces DevMentorAI, an LLM-based tool that uses a RAG approach to answer developer questions, which is then evaluated with a case study on Mozilla's PDF.js proje

Strengthening Data Ecosystems in Indian Schools

Educational data collection and analysis can both improve and introduce risks to India’s sprawling education system. Findings follow a nine-month investigation by Mozilla and Aapti Institute, supported by USAID.

Towards a Framework for Openness in Foundation Models

This paper presents a framework for grappling with openness across the AI stack. The paper surveys existing approaches to defining openness in AI models and systems, and then proposes a descriptive framework to understand how each component of the foundation model stack contributes to openness.

Mind the Gap: What Working With Developers on Fuzz Tests Taught Us About Coverage Gaps

AI Intersections Database

The AI Intersections Database maps intersections between the key social justice and human rights areas of our time and documented AI impacts and their manifestations in society.

Full Disclosure: Stress testing tech platforms’ ad repositories

Mozilla called on Check First to test the ad transparency tools maintained by 11 of the world's largest tech companies. These tools aim to let researchers, watchdogs, and members of the public understand how commercial communications influence the information space and affect society.