Integrating Machine-Generated Short Descriptions into the Wikipedia Android App: A Pilot Deployment of Descartes

2026-01-12

AI Summary

This study addresses the uneven coverage of Wikipedia’s short descriptions across languages and topics, which adversely affects user experience. The authors present the first large-scale deployment of the multilingual generative model Descartes within the Wikipedia Android application, providing real-time machine-generated short description suggestions to editors. Integrated community feedback mechanisms were employed to evaluate the usability and safety of these suggestions. The intervention spanned 12 languages and over 3,900 articles, with 90% of adopted generated descriptions receiving quality scores of at least 3 out of 5—comparable to human-written counterparts. Furthermore, low edit revision and revert rates demonstrate the method’s effectiveness and scalability in authentic editing environments.

Technology Category

• Natural Language Processing: Generation
• Machine Learning: Large Multimodal Models (LMMs)
• Application Domains: Humanities & Computational Social Science

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Web Mining and Content Analysis: Web data quality in the era of algorithmically-generated content
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work

Abstract

Short descriptions are a key part of the Wikipedia user experience, but their coverage remains uneven across languages and topics. In previous work, we introduced Descartes, a multilingual model for generating short descriptions. In this report, we present the results of a pilot deployment of Descartes in the Wikipedia Android app, where editors were offered suggestions based on outputs from Descartes while editing short descriptions. The experiment spanned 12 languages, with over 3,900 articles and 375 editors participating. Overall, 90% of accepted Descartes descriptions were rated at least 3 out of 5 in quality, and their average ratings were comparable to human-written ones. Editors adopted machine suggestions both directly and with modifications, while the rate of reverts and reports remained low. The pilot also revealed practical considerations for deployment, including latency, language-specific gaps, and the need for safeguards around sensitive topics. These results indicate that Descartes's short descriptions can support editors in reducing content gaps, provided that technical, design, and community guardrails are in place.

Problem

• short descriptions
• content gaps
• Wikipedia
• multilingual coverage
• user experience

Innovation

• machine-generated descriptions
• multilingual AI
• human-AI collaboration
• Wikipedia editing
• pilot deployment