Human knowledge at stake: Wikimedia Foundation makes its case at the United Nations General Assembly

Speaking at the Clinton Global Initiative main stage, Meehan framed the challenge of protecting human knowledge as a defining one for the AI era: "Our moonshot for this era is to keep humans at the center of this evolving knowledge ecosystem," she said."It is not enough to develop and design these [AI] systems; we have to keep humans not only in the center of them, but at the heart of them." The panel spoke to the value of human knowledge online:

• Meehan described free knowledge projects like Wikipedia as providing the cumulative memory of all of humanity - open to everyone to access and contribute, and belonging to all of us..
• Generative AI cannot survive without human knowledge. Wikipedia is trusted and used by AI models because it is built transparently by real people, and edited in public, with standards that promise verifiability, neutrality, and transparency.
• Protecting digital public spaces like Wikipedia helps keep human knowledge and human voices at the heart of the internet.

At the UN's Sustainable Development Goals Media Zone, Meehan was candid about the cost of providing free information. "Wikipedia's content is free. We don't run ads. We don't monetize. But our infrastructure is not free, and the toll of scraping and the behavior of some of the big technology and frontier AI companies takes a toll on our infrastructure and how we're able to support our free knowledge projects. Together, the panelists explored what the toll looks like in practice, and what companies who rely on sources like Wikipedia should do to use them sustainably:

• Meehan highlighted that rising bot traffic and scraping is taking a real toll on the Wikimedia projects. It requires more and more personnel and resources to maintain the infrastructure that makes Wikipedia and other Wikimedia projects freely available to millions of people around the world.
• She described how an AI-dominated internet has made digital public goods like Wikipedia more vital, but less visible, meaning that they may struggle in the future to attract more participation and support. Wikipedia requires a constant flow of volunteer contributions and financial support from its users. When AI platforms and search engines provide quick machine-generated summaries and responses to users without attributing it back to Wikipedia, they block pathways for people to access the original sources of information to fact-check, research further, and contribute to Wikipedia.
• She highlighted the need for large-scale commercial content reusers to recognize that the value of their products depends on Wikipedia's human-created knowledge, which supports the wider information ecosystem. Without Wikipedia, AI models will lack the diverse, verifiable, and open data that fuels innovation in ways that proprietary content alone cannot. They need to support the long-term sustainability of Wikipedia, which they rely so heavily on: through content attribution and financial support.

During the week, UNGA attendees across events discussed the recent headlines centered on AI safety, as well as the need to look more widely at the challenges being posed by these new technologies. In her talk on Devex Impact House "The AI governance discussion today is a lot like a dragon. Everyone is focused on the teeth and the fire coming out of the dragon's mouth-the frontier labs and how AI could spiral out of control." said Meehan."But while that conversation is important, we also need to worry about the dragon's tail destroying the buildings behind it. We need to talk about the damage being done to the very foundation of the online information ecosystem." As the week came to a close, discussions focused on why conversations around AI governance and regulation need a wider lens, and who should be included in those conversations to ensure we're protecting the human knowledge sources AI depends on in the first place. Key points that were raised include:

• AI safety is critical, but if we lose the information ecosystem that feeds the technology in the first place, we are missing the point. To regulate properly, we need to be talking about the AI inputs as well as the outputs.
• As lawmakers across the globe grapple with how to regulate AI and digital platforms, it's critical that they don't inadvertently harm the digital public goods the technology depends on. The Wikipedia test can help ensure that new regulation preserves the spaces and parts of the internet we love the most.
• Civil society, multilateral organizations, academia, the media ecosystem, and experts alike need to be at the table to ensure the policy conversations we're having benefit all of us - not just a handful of companies.

Wikimedia_UNGA_IMG_3281

Lucia Liss

Wikimedia Foundation Logo

Wikimedia Foundation

Rena Kokalari

Image by Rena Kokalari

Jens Mohr