The Wikimedia Foundation has suspended a test of displaying summaries above articles using artificial intelligence (AI) that was announced on June 2 after receiving criticism from editors, according to a 404 Media report. The feature was supposed to be tested for two weeks and only appeared on the mobile version.
Wikipedia's editorial community rejected the introduction of the feature because it harmed the site's reputation as a reliable source. Like Google's AI Overviews, the feature that Wikipedia is testing would summarize long articles on the site into short, more understandable paragraphs.
With Google facing controversy over inaccurate AI Overviews, Wikipedia editors feel that the same thing could happen, causing readers to receive incorrect information. The editors also stated that Wikipedia content is long and boring to read, but this should not be considered a negative, especially for complex information.
Although the AI summary test has been discontinued by Wikimedia, they are still interested in continuing to use it to make the site's content more accessible to readers around the world. The Wikimedia Foundation uses Cohere's Aya model to perform summarization and has previously agreed to license the content for AI training purposes as well as using AI to translate English articles into other languages.