Earlier this week we saw the announcement of an extremely interesting alliance to promote the development and deployment of open AI called the AI Alliance. The group includes not only a number of leading industry players, but also a group of university partners and government agencies (including the US National Science Foundation and NASA); the group is also highly international and includes organizations like CERN and ETH.

For the announcement and additional information see

https://thealliance.ai/news

It's always difficult to predict what long-term impact these kinds of alliances will have, but this feels timely and significant, and is a welcome counterbalance to trends that have been promoting centralized and rather opaque large foundation models and systems built upon them.

It's worth noting that the notion of "open" AI is quite complex and certainly not well understood or generally agreed upon. Because of the way code, data and processes (perhaps including human interactions) are intertwined in the construction of AI systems it's clear that the traditional understandings based on open source are insufficient. For some discussion of this see

https://deepdive.opensource.org/wp-content/uploads/2023/02/Deep-Dive-AI-final-report.pdf

and a very useful paper "Open (For Business): Big Tech, Concentrated Power, and the Political Economy of Open AI" (August 2023) at

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4543807


Clifford Lynch
Director, CNI