Français
Most open source AI conversations are about code, weights, and compute… Almost none of them are about law. That’s the gap Vincent Bachelet works in.
A postdoctoral fellow at Université Paris 1 Panthéon-Sorbonne and legal consultant at Inno3 SAS, Vincent focuses on the legal dimensions of open source AI strategy.
At POSAIS 2026, Vincent raised a question with much bigger implications: AI doesn’t map cleanly onto the legal frameworks that made open source software work in the first place.
Weights, training data, licensing, responsibility, the legal questions become much more complex once you look closely. “We’re still able to build a legal framework for true open source AI.”
What Vincent values about events like POSAIS is precisely this: bringing technical, legal and strategic perspectives into the conversation from the start, rather than bringing in legal expertise once the model already exists.
It’s the part of “open source AI” that rarely makes a keynote slide.
And often, it’s the part that determines whether the whole thing can actually be shared, used and built upon.
