Wikidata

Gold definitionUpdated Apr 2, 2026

Wikidata is a central, open, and collaborative knowledge base hosted by the Wikimedia Foundation, launched in 2012. It serves as a structured data repository for information across all Wikimedia projects, including Wikipedia, and is freely available for use by anyone. At its core, Wikidata organizes information as items (representing entities like people, concepts, or events) and statements (key-value pairs describing properties of these items, linked to other items or literal values). This structured approach allows machines to easily read, query, and process data, enabling semantic web applications, data integration, and knowledge graph construction. It addresses the challenge of inconsistent data across different language Wikipedias and provides a unified source of factual information, crucial for applications ranging from search engines and voice assistants to advanced AI systems that require robust, verifiable knowledge. Researchers and developers widely use Wikidata as a foundational knowledge graph for tasks like entity linking, question answering, and enhancing large language models.

Key Aspects of Wikidata

Items and Properties: Wikidata organizes information around 'items,' each representing a concept, entity, or topic, identified by a unique Q-ID. Properties (P-IDs) describe relationships or attributes of these items, linking them to other items or data values.
Structured Data and Linked Open Data: All data in Wikidata is structured, making it machine-readable and queryable via SPARQL. It adheres to Linked Open Data principles, allowing seamless integration with other datasets across the web, fostering a vast network of interconnected knowledge.
Multilingual Support: Wikidata is inherently multilingual, allowing labels, descriptions, and aliases for items and properties in hundreds of languages. This enables cross-lingual data access and supports global knowledge sharing without language barriers.

Applications of Wikidata in AI/ML

At a glance

Executive summary

Wikidata is a free, open, and collaborative knowledge base that stores structured data for Wikipedia and other projects. It makes factual information machine-readable and queryable, supporting AI applications that need reliable, structured knowledge.

TL;DR

Wikidata is a giant, free, online database of facts organized so computers can easily understand and use the information.

Key points

Stores factual information as structured data (items and properties) in a collaborative, multilingual knowledge base.
Provides a unified, machine-readable source of verifiable factual knowledge, addressing data inconsistencies and enabling semantic web applications.
Used by Wikimedia projects, search engines, voice assistants, AI researchers, and developers for LLM grounding, entity linking, and knowledge graph applications.
Unlike unstructured text (e.g., Wikipedia articles), Wikidata offers explicit, machine-readable relationships and types, enabling precise querying and reasoning.
Increasingly used as a foundational knowledge graph for grounding Large Language Models, improving their factual accuracy and reducing hallucinations.

Use cases

Powering Wikipedia Infoboxes: Automatically populating factual summaries in Wikipedia articles across different languages.
Google Search Knowledge Panels: Providing structured data for the information boxes that appear on the right side of Google search results.
Voice Assistants (e.g., Siri, Alexa): Answering factual questions by querying its structured data, such as 'Who is the current president of France?'.
LLM Factual Grounding: Supplying verifiable facts to Large Language Models to prevent them from generating incorrect or fabricated information.
Academic Research: Used in computational linguistics, data science, and AI for tasks like entity recognition, relation extraction, and building specialized knowledge graphs.

Also known as

Q-ID, P-ID, Wikimedia Knowledge Base, Linked Data