Episode 77 •
24 July 2024
Elizabeth Seger on Open Source AI
Contents
Elizabeth Seger is the Director of Technology Policy at Demos, a cross-party UK think tank with a program on trustworthy AI. She was previously a researcher with Centre for the Governance of AI (GovAI) in Oxford, and is an affiliate of the AI: Futures and Responsibility Project (AI:FAR) at the University of Cambridge.
In this episode we talked about open source the risks and benefits of open source AI models. In particular, we discussed:
- What ‘open source’ really means
- What is (and isn’t) open about ‘open source’ AI models
- How open source weights and code are useful for AI safety research
- How and when the costs of open sourcing frontier model weights might outweigh the benefits
- Analogies to ‘open sourcing nuclear designs’ and the open science movement
Resources
- Report: Open-sourcing highly capable foundation models (Seger et al.)
- Lawrence Lessig on open source AI risks
- Podcast: The Promise and Peril of Open Source AI with Elizabeth Seger and Jeffrey Ladish
- Dependency (XKCD comic)
- Mozilla’s new framework for AI openness and innovation
- Meta’s AI language model LLaMA leaked online
- The tech industry’s open source AI definition problem
- New York Times: What to know about open vs. closed software
- Stanford CRFM paper on open foundation models
- Open source AI: A regulatory review
- Vox article on open source AI risks and benefits
- Ajeya Cotra on open source AI
- The definitive story of “Information wants to be free”
- UK Government’s INSPECT AI project
- NTIA receives over 300 comments on open-weight AI models
- Mozilla Foundation: Technical readout of Columbia convening on openness and AI
Let us know if we missed any resources and we’ll add them.
Transcript
Coming soon!
Note that this transcript will be machine-generated, by a model which makes frequent mistakes and sometimes hallucinates entire sentences. Please check with the original audio before using this transcript to quote our guest.