In the AI era, Wikipedia has never been more valuable (wikimediafoundation.org)

As generative AI floods the web with synthesized text and media, the Wikimedia Foundation argues that Wikipedia’s human-curated, multilingual corpus has never been more essential. Volunteer editors—hundreds of thousands strong—provide the discussion, sourcing, and ongoing updates that AI cannot replicate: discovering underdocumented facts, enforcing neutrality, and publicly logging editorial debates and citations. That human process underpins verifiability and helps prevent AI “hallucinations” or long-term model degradation that can occur if systems lose access to continually refreshed, vetted knowledge. For the AI/ML community this has practical implications: Wikipedia is one of the highest-quality datasets for training and evaluation, and omitting it degrades accuracy, diversity, and verifiability in model outputs. Wikimedia asks developers to treat its content responsibly by attributing sources and financially supporting sustainable access—principally via Wikimedia Enterprise, a paid opt-in product that reduces server strain while funding the project. The Foundation also embraces limited, volunteer-governed AI tools to automate mundane tasks (e.g., vandalism flagging) so editors can focus on substantive work. In short, maintaining Wikipedia’s open, human-led ecosystem is framed as a technical and ethical imperative for building reliable, long-lived AI systems.