Mozilla Automation TeamMark Surman3 min readintro
A New Framework for Open Source AI
Summary
Mozilla and partners published a paper proposing a layered, gradient openness framework for foundation models, defining openness for data, code, weights, docs, and deployment. The framework gives developers, regulators, and civil society a common language to evaluate openness and safety beyond a binary label.
- The framework breaks AI systems into layers (data, code, weights, documentation, deployment) and defines openness per layer.
- Openness is modeled as a gradient, allowing mixed states (e.g., open weights but closed training data).
- Safety must be evaluated across the entire stack, not just the model weights.
- Provides a shared vocabulary for developers, researchers, and regulators to assess openness trade‑offs case‑by‑case.
Developers, policymakers, and researchers building or regulating foundation models need a precise vocabulary to reason about openness and its safety implications.
5/10
