Jimmy Wales, the founder of Wikipedia, has suggested that artificial intelligence could be used to improve how the online encyclopedia handles draft articles. His comments, posted on his Wikipedia user talk page in July, have reignited a long-standing debate within the community about whether AI tools align with the values that underpin Wikipedia. Wales views AI as a possible way to provide better guidance to new contributors, but many experienced editors have pushed back strongly, arguing that introducing AI into the editing process could undermine the reliability and reputation of the site. The exchange highlights a growing divide over how new technologies should interact with one of the internet’s largest volunteer-driven knowledge projects.
Wales’ Proposal to Use AI in Draft Review
Wikipedia’s draft process allows anyone to submit an article, but before it can appear in the main encyclopedia, the draft must be reviewed by experienced editors. Because hundreds of submissions arrive daily, reviewers often rely on a tool known as the Articles for Creation Helper Script (AFCH), which provides templated reasons for rejecting drafts. While efficient, this system often frustrates new contributors who receive only a generic explanation without detailed feedback. Wales argued that this approach can discourage contributors from improving their work, since they may not fully understand what needs to be fixed.
In one example, Wales described being asked to evaluate a draft article that had been repeatedly rejected using template responses. Feeling that the existing feedback was unhelpful, he decided to try an experiment. He submitted the text to ChatGPT and asked the system for suggestions on how to improve it. The AI-generated response, he said, contained more specific guidance than the standard templates and might offer new editors a clearer path forward. He suggested that incorporating AI into the AFCH system could make the review process more constructive.
“It would be better, obviously, if we had lovingly crafted human responses to every situation like this, but we all know that the volunteers who are dealing with a high volume of various situations cannot reasonably have time to do it,” Wales wrote. “The templates are helpful, and an AI-written note could be even more helpful.” His comments emphasized that AI would not replace human reviewers but could act as a support tool to reduce frustration and improve retention among new contributors.
Editors Raise Concerns Over Accuracy and Reliability
Many veteran Wikipedia editors were quick to point out problems with Wales’ proposal. They noted that the ChatGPT feedback he shared contained factual inaccuracies and improper recommendations, such as citing press releases from Harvard Business School as sources. Wikipedia policy explicitly bars press releases from being used to establish notability because they are considered non-independent. Editors argued that new users might mistakenly believe the AI’s advice was accurate and begin replicating these mistakes throughout the site, leading to a wider decline in quality.
The critics also raised concerns about AI’s inability to fully grasp core Wikipedia principles, such as notability, verifiability, and neutrality. These policies are central to ensuring that articles reflect reliable sources and balanced perspectives. Editors said that even small inaccuracies in AI-generated guidance could distort how contributors interpret these rules, potentially causing widespread issues if newcomers internalize faulty advice. “If they believe the LLM advice accurately reflects our policies, and it is wrong or inaccurate even 5 percent of the time, they will learn a skewed version of our policies,” one editor warned.
The most forceful objections highlighted reputational risks. Editors argued that allowing AI-generated content into Wikipedia’s editorial process could damage trust in the platform. One response described the proposal as “the antithesis of Wikipedia,” stressing that the project thrives on human cooperation, mentorship, and consensus-building. Automated feedback, even if well-intentioned, was described as an illusion of helpfulness that discourages true learning and undermines the volunteer-driven standards that have sustained Wikipedia since its creation.
Ongoing Debate Over AI’s Role in Wikipedia
The disagreement between Wales and editors is part of a broader conversation about AI’s place in the future of Wikipedia. Earlier this year, the Wikimedia Foundation, the non-profit that operates the platform, paused an experiment involving AI-generated summaries at the top of articles after significant pushback from editors. In August, the community also approved a new policy that makes it easier to delete content that is clearly AI-generated, signaling a strong commitment to preserving human-led editorial work. These steps reflect a cautious stance toward automation, even as AI tools become more common across the internet.
At the same time, the Wikimedia Foundation has acknowledged that AI and automation already have a role in maintaining the site. For years, bots have been used to detect and revert vandalism, and machine translation tools have helped contributors work across languages. Importantly, these systems always require human oversight. A spokesperson emphasized that any AI use on the platform must follow the principle of having “a human in the loop,” ensuring that automated work is reviewed and validated by volunteers before it becomes part of the encyclopedia.
In a statement, the Foundation underscored the importance of keeping Wikipedia grounded in its human-centered model. “Wikipedia’s strength has been and always will be its human-centered, volunteer-driven model, one where knowledge is created and reviewed by people, volunteers from different countries, perspectives, and backgrounds,” a spokesperson said. At the same time, the spokesperson noted that machine-generated content is “exploding across the internet” and will inevitably appear on Wikipedia in some form. Wales’ comments, they added, were an example of his regular engagement with the community to test ideas and spark discussion, rather than a formal directive.