Wiki Education's 2025 analysis reveals the impact of generative AI on Wikipedia editing. They discovered participants were increasingly using AI tools like ChatGPT, often resulting in inaccurate content. The study examined edits from their programs, which contribute significantly to new editor activity on the platform, and found a concerning rise in the use of AI-generated text.

The investigation utilized the Pangram AI detection tool, which flagged articles created with AI. The results showed a correlation between the launch of ChatGPT and an increase in articles with AI-generated content. A major issue was the failure of verification, where information cited to real sources did not exist in those sources. This led to significant time spent on cleanup.

To address this, Wiki Education implemented new training modules and real-time AI detection using Pangram. They provided guidance to participants on the appropriate use of AI tools, emphasizing the prohibition of directly copying and pasting AI chatbot outputs. The goal was to maintain the accuracy and integrity of Wikipedia content, a critical challenge in the age of widespread generative AI.

Looking ahead, the Wikimedia Foundation and other organizations are grappling with how to manage AI-generated content on collaborative platforms. The ongoing challenge is balancing the potential benefits of AI tools with the need to ensure factual accuracy and prevent the spread of misinformation. It's a key area to watch in the coming years.