Attribution, Commons and Open Source AI: OKFN in Paris for the DPGA Annual Members Meeting

This month we are in Paris for the Digital Public Goods Alliance Annual Members Meeting, hosted by the Government of France. The AMM is where the DPGA community comes together every year to discover, develop, use and invest in digital public goods. Our delegation will be there throughout, leading sessions concentrated on 18 November.

Recent research found that more than 30% of responses from leading consumer AI tools carried no attribution at all, and where sources were cited, they were inaccurate 72% of the time — missing, unsupported or unverifiable. In Enabling Attribution for AI Systems through DPGA Standard Indicators, we join Creative Commons, Open Future, the Open Source Initiative and the DPGA to ask how the DPG Standard could make attribution the norm rather than an afterthought. We will share what we learned from our MCP pilot with the governments of Brazil and Uruguay (also DPGA members), where the aim was to go beyond attribution towards provenance: every answer traceable back to the dataset it came from.

Two other sessions follow. Shielding the Commons: Policy Responses to Urgent Risks and AI Extraction, with the Wikimedia Foundation, is an interactive workshop built around a policy framework that treats open knowledge systems as critical digital infrastructure — participants map their own role in those systems, then work through realistic national crisis scenarios to test what holds.

Later, with the Open Source Initiative, Creative Commons and the DPGA Standards Team, our team will get into the practical questions that have lingered since the Open Source AI Definition was released. The session Practical Discussions on Implementation of the Open Source AI Definition will discuss licensing, data privacy, compliance, verification, and whether distinguishing open weights from Open Source AI actually helps.

None of these topics are abstract debates for us. Attribution, provenance and the governance of shared resources are the same questions we keep running into across our AI Learning Labs, our Frontiers of Openness work and The Sharing Shift. Being in the room with the people building standards, running national infrastructure and drafting policy is how those questions move from experiment to norm.

If you’re in Paris for the AMM, come and find us — our team would be glad to connect. You can also reach us at info@okfn.org, or follow our social media channels (LinkedIn, Bluesky, Mastodon, X) for updates from the sessions.