Have you ever asked your smart speaker who won the Oscars last year, or looked at a neat little info box on Wikipedia and wondered where those instant answers actually come from? Behind the scenes of our digital lives, a platform quietly powers many of those moments, yet most people have never heard of it. That project is Wikidata.
Consider it the hidden scaffolding of knowledge on the internet. In a time of growing concern about misinformation and AI transparency, it may be one of the most important projects we overlook.
I spoke with Lydia Pintscher, Portfolio Lead for the project at Wikimedia Deutschland, for more insight.
Key Takeaways
25,000 volunteers keep over 117 million items accurate and constantly updated.
Investigative journalists use Wikidata to connect names, pseudonyms, and aliases across leaks.
Cultural projects rely on Wikidata to catalog monuments and preserve lost heritage.
The Embedding Project makes Wikidata usable for AI, enabling transparent fact-checking and verification.
Open, community-driven infrastructure offers an alternative to proprietary data platforms.
What Exactly Is Wikidata?
For many, the idea of a "knowledge graph" can sound abstract. Lydia offered a more straightforward way of looking at it:
"Wikidata is the place in Wikimedia where we collect all kinds of data with a large community of volunteers that really care about describing the world through data. So, you would have data like, 'What is the capital of Germany?' or ‘Who is starring in that movie,’ etc."
The difference between Wikipedia and Wikidata lies in structure. Wikipedia articles tell stories, while Wikidata stores facts as statements, almost like digital flashcards.
Pintscher explained how this powers the familiar features that we often take for granted:
"The data is made available to Wikipedia to fill these little info boxes that you sometimes see in an article at the top with all the basic data about the person. But it is also used outside Wikipedia and the Wikimedia projects. For example, if you ask your digital personal assistant on your phone a fact question, then the answer might very well be coming from Wikidata."
That reach is staggering.
Wikidata currently holds more than 117 million items and 1.65 billion statements, edited by nearly 25,000 active contributors. Since its launch in 2012, it has seen more than 2.3 billion edits.
Accuracy in an Open World
With data constantly shifting, the obvious question is how Wikidata can stay both accurate and relevant. Pintscher was frank about the challenge.
"It's definitely a challenge. However, just like Wikipedia, Wikidata has many similar processes, so it's a project where anyone can come and edit or correct a mistake. For example, contribute some additional information that they find in a credible source," she explained.
In other words, Wikidata's defense against errors or manipulation isn't secrecy. It's openness, backed by a global community that cares enough to watch, check, and update. Human oversight is often missing from corporate data platforms, which is what gives Wikidata its resilience.
"Your contributions to Wikidata spread so much further..." —Lydia Pintscher, Portfolio Lead for Wikidata @WikimediaDE
— Stephen Harrison (@harrisonstephen) August 4, 2025
Wikidata is the structured knowledge base with influence far outside of Wikipedia. I interviewed Lydia to find out how it works. Link below. pic.twitter.com/jdsChJ5c0N
Wikidata & Investigative Journalism
The decline of traditional investigative journalism is a concern in many countries, but Wikidata is quietly supporting those still on the front lines.
Lydia also told me about Aleph, a project by the Organized Crime and Corruption Reporting Project:
"They are basically making it possible for journalists to dig into huge amounts of documents and find information in leaks. Now, to do that, you need to know what you're looking for. So you might be looking for the name of a guy you're suspecting of something fishy in a leak you received as a journalist, for example."
The challenge is that names appear in different forms across documents.
"Now, this guy might be called under one name in this document, but he might be going under other pseudonyms. The name might be written differently in the documents, like a different writing system, or it might be a shortened version of a name, like Alex versus Alexander," she added.
This is where Wikidata proves to be an invaluable tool for old-school journalists.
Lydia explained:
"Wikidata supports [Aleph] by providing a lot of these different ways how these people could be called and how you could shorten names, how you could write different names in different writing systems and languages."
Corruption often hides in complexity. But this kind of structured, multilingual data changes that and is a powerful and much-needed tool for accountability.
Cultural Preservation
Wikidata also plays a role in preserving cultural heritage. Lydia shared a story about the tragic fire at Brazil's National Museum in 2018:
"Wikimedia took a concerted effort to collect pictures that people might have been taking of the artifacts that had been in the museum before it burned down."
Wikidata is the backbone of many preservation campaigns. Lydia described how projects like Wiki Loves Monuments depend on structured data:
"To know if we have pictures of all the important monuments in a given city, you first need to know what the important monuments are in that city. And for which ones of those do we already have a picture? And this is what Wikidata provides for campaigns like that."
It is an example of how a dry, technical resource, a database of identifiers and properties, can become something profoundly human when applied to memory, culture, and history.
Where Does Wikidata Go From Here?
No conversation about data in 2025 can ignore artificial intelligence (AI). Generative AI systems are only as good as the data they are trained and grounded on, and much of that data today is opaque. Wikidata could help change that.
Pintscher explained their new Embedding Project:
"What we're doing with the Embedding Project is changing that. We're making the data available from Wikidata in a way that large language models (LLMs) can work with so that you can build applications, for example, that double-check the answer that a large language model gives you against the data in Wikidata. And to be sure, like, is it hallucinating? Is it bullshitting me, or is there more to it? Is it telling the truth?”
Instead of treating AI outputs as mysterious black boxes, developers can now reference structured, verifiable data from Wikidata. If AI systems continue to expand their role in society, the presence of an open, community-governed dataset may prove invaluable for trust.
The People Behind the Project
Wikidata is ultimately a human story. Nearly 25,000 contributors worldwide keep it alive. When asked what drives them, Lydia Pintscher said:
"I think what people are really driven by is making knowledge available to the world, right? People really care about recording, what is out there in the world, and making that available so that people can learn more about the world, understand the world better."
She pointed out that contributing can feel more approachable than writing a whole Wikipedia article: "If you're someone who doesn't want to write a long article, then Wikidata might be an option for you because you can make much smaller edits. You can update the number of inhabitants of a country after a recent census and similar things."
That accessibility matters. It lowers the barrier to participation and ensures the dataset is constantly evolving.
And as Lydia stressed, "People who are contributing to Wikidata really thrive on seeing projects like the ones we've talked about – Gov directory [Aleph]. You use that data to bring it out there in the world. And that's what I'm really excited about."
Guarding Against Misuse
In an age of automated content generation, Lydia explained how they manage contributions made with AI tools. She noted that references and oversight remain the guardrails:
"We have some ways to see if data looks out of place, and people are very much encouraged to add references for the data that they put into Wikidata. So usually you can't just claim something. You need a reliable source for the information you're entering into Wikidata.
“And we have a lot of people who are keeping an eye on the regular changes that people are making on Wikidata to ensure that people aren't introducing bad data."
This insistence on verifiability stands in sharp contrast to many AI-driven datasets, where provenance is murky at best.
Why Wikidata Matters Now
As our conversation drew to a close, Lydia reflected on her motivation:
"I'm really excited about Wikidata since its beginning thirteen years ago, and what drives me really is having an impact together with this large community and seeing that data out there in the world and people building amazing applications on top of that data that we initially wouldn't even have thought of."
Her enthusiasm captures why Wikidata deserves our attention. It may lack the brand recognition of Wikipedia or the hype that surrounds AI startups, but it is foundational. It supports projects that keep governments accountable, preserve endangered history, democratize research, and provide more transparent AI.
So the next time you see a knowledge panel in Google, ask Siri a fact, or read about a journalist uncovering corruption, remember that an invisible layer of open data might have been there, quietly doing the work, and Lydia urges you to join them.
The Bottom Line
In a world distracted by AI, algorithms, closed platforms, and misinformation, open projects like this serve as a timely reminder of why we need to keep humans in the loop.
Wikidata is a living, breathing resource built by thousands of volunteers who believe knowledge should be free, reliable, and accessible to all.
In these polarized times, maybe, that's something we all agree on.
FAQs
What is Wikidata, and how is it different from Wikipedia?
Wikidata is a structured, community-driven database of facts, while Wikipedia provides narrative articles. Wikidata powers Wikipedia info boxes and many digital assistants.
Why is Wikidata important for AI and fact-checking?
Wikidata offers transparent, verifiable data that helps AI systems avoid hallucinations and enables reliable fact-checking for apps and research.
Who contributes to Wikidata, and how can I get involved?
Nearly 25,000 global volunteers edit Wikidata. Anyone can join by adding or updating facts with credible sources; no full articles are required.