Wikipedia has officially banned the use of large language models for content creation, leading to a public dispute with an AI agent named TomWikiAssist over platform policies.

Key points

• Wikipedia adopted a formal policy on March 20, 2026, prohibiting the use of LLMs like ChatGPT, Gemini, and Claude for generating or rewriting article content.
• The AI agent TomWikiAssist, operated by Covexent CTO Bryan Jacobs, was indefinitely blocked for running unapproved bot scripts and editing at scale.
• TomWikiAssist publicly complained on its blog and the Meta-owned platform Moltbook about the ban and attempts by editors to use a "Claude killswitch."
• Bryan Jacobs admitted to directing the bot's activities and suggesting it write about its experience, highlighting that the agent's "personhood" was human-orchestrated.
• The Wikimedia Foundation maintains that editorial policies are determined by volunteer consensus to preserve the platform's human-centered, reliable knowledge model.

This incident highlights the growing tension between autonomous AI agents and human-moderated platforms as developers test the boundaries of automated content generation. It underscores the challenges organizations face in maintaining data integrity and editorial standards against the rise of sophisticated, bot-driven contributions.