Wikipedia and AI: Why Academics Should Defend It
Generative AI relies on Wikipedia while diverting its readers. See why academics should defend open knowledge and how universities can support them in practice.
If AI Depends on Wikipedia, It Has a Stake in Keeping It Alive
Generative AI can answer a question in seconds. For a lot of people, that's the whole decision: they read the response, skip the sources, and move on. Convenience is real - but there's a quieter downside. If fewer people click through to the underlying knowledge sites, including Wikipedia, those sites get less support over time.
That's a practical issue, not just a philosophical debate about whether AI answers are "good enough." Wikipedia relies on people who read, edit, discuss, and maintain it. If visits drop and fewer readers become contributors, the volunteer community may have less capacity to update articles, fix errors, and fill gaps. Meanwhile, AI companies can draw on Wikipedia's accumulated knowledge to build products that compete for the same attention.
Academics are tied to this in a few roles: researchers, teachers, and participants in how knowledge gets produced. That doesn't mean every scholar needs to become a Wikipedia editor. But it does raise a more useful question: what can academics and universities reasonably do to help sustain a public knowledge resource their students and institutions rely on?
This article looks at the pressure generative AI puts on open knowledge, why academic involvement is ethically defensible, and practical ways to contribute without treating Wikipedia as a substitute for scholarly research.
What generative AI changes for Wikipedia
Wikipedia has always had to compete for attention. Search engines, social platforms, and commercial reference products all aim for the same users. Generative AI shifts the interaction: it can deliver a polished answer that sits between a person and the sources that helped produce it.
If someone asks ChatGPT or Google Gemini a factual question, they may get a direct response, sometimes with citations or links. But a link isn't the same as a visit. If a user accepts the generated answer without opening the underlying material, Wikipedia may contribute information without gaining a reader - someone who might notice a problem, correct it, or share the article with others.
The reader-to-contributor loop
Wikipedia isn't run like a traditional newsroom with staff writers responsible for every page. It depends on a distributed volunteer community, and readers are part of the pipeline. When people spot an error, they can add a citation, improve an explanation, or take on more sustained editing.
That relationship works like this:
- People use Wikipedia to find information.
- Some readers contribute corrections, sources, or new material.
- Those edits improve the resource for later readers.
- The community reviews and maintains the work.
If AI tools reduce visits, they can weaken that route into the cycle. This doesn't mean every lost click becomes a lost editor, or that traffic alone decides Wikipedia's future. Still, changes in how people discover answers can change the conditions that keep volunteer knowledge projects going.
A resource and an input
There's also a mismatch worth naming. Wikipedia is a public reference work, but its content is useful to companies building AI products. Some companies access Wikipedia through paid, high-speed API services; others may scrape. The details differ, so it's not accurate to treat every use as the same. The underlying tension is consistent, though: knowledge created and maintained by volunteers can end up powering commercial products that compete for the same users.
Wikipedia is more than a dataset. It's a community with norms, editorial processes, and a degree of independence from commercial or government control. Treating it as just machine-readable text misses the work required to keep knowledge current, sourced, and genuinely contested when needed.
The concern isn't "AI will use Wikipedia" by itself. Reuse can be legitimate and even helpful. The question is whether the systems that benefit from open knowledge also support the conditions that let that knowledge exist in the first place. A generated answer can hide uncertainty, skip context, or repeat an error. When that happens, users need access to sources - and institutions or communities that can correct the record. Open knowledge needs readers, contributors, and transparent editorial work, not just a place where facts get absorbed into models.
Why academics have a particular responsibility
The ethical case for academic involvement isn't that universities created the problem, or that scholars alone can fix it. It's that academics are closely connected to the knowledge systems affected by it - and they often have skills and influence that matter in practice.
Helen Choi makes this case through Iris Marion Young's Social Connection Model. The basic idea is that responsibility can arise from participation in the structures and processes that produce injustice. Applied here, the argument is straightforward: people and institutions tied to knowledge production have reason to pay attention to how these systems work, and who absorbs the costs when they fail.
Academics research, evaluate sources, teach students, and shape what counts as credible knowledge. They also work inside institutions that typically have more resources and formal authority than most individual Wikipedia volunteers. That combination creates meaningful capacity - not unlimited responsibility, but more than the ability to watch from the sidelines.
Responsibility is not the same as blame
This distinction matters. One lecturer isn't responsible for every detail of how every commercial AI product is designed, and no single person can "preserve" a global encyclopedia. Turning a structural problem into a demand that overextended academics volunteer indefinitely would be a poor solution.
A better approach is to align responsibility with connection and capacity. Academics can contribute subject expertise, teach information literacy, model transparent citation practices, and advocate for institutional recognition. Universities can offer time, training, and support. At the same time, AI companies and platforms should be accountable for how they use knowledge resources and for how their products affect access to the sources that let people verify and correct information.
This is shared responsibility, not a one-way transfer of work from companies to universities and volunteers.
Expertise can improve public understanding
Academic knowledge helps Wikipedia when it supports clear, reliable, well-sourced explanations that readers can actually use. Specialists can spot where an article lacks context, where a claim leans on weak sourcing, or where a technical explanation is likely to mislead someone outside the field.
That doesn't mean Wikipedia becomes a scholarly journal. Its audience, purpose, and editorial model are different. Academic credentials don't override Wikipedia's community rules, and expertise in a topic doesn't automatically make every edit appropriate. The strongest contributions usually come from pairing subject knowledge with solid independent sources, neutral language, and respect for collaborative review.
The same lesson applies to AI. Fluent output isn't evidence that a claim is correct. Researchers who teach students to trace claims back to sources are building a habit that matters regardless of whether the starting point was a chatbot, a search result, a textbook, or Wikipedia itself.
How academics can support Wikipedia in practice
There isn't one contribution model that fits everyone. A reasonable starting point is to pick an action that matches your expertise, your time, and your role. Even small, well-sourced improvements can matter more than a large editing project that falls apart after a week.
1. Improve an article within your area of expertise
Start with a topic you know well, and look at the article before you edit. Identify real issues: an outdated statistic, missing context, a weak or broken citation, an important omission, or wording that makes a contested claim sound more certain than the sources actually support.
Then work within Wikipedia's expectations. Use high-quality published sources, make claims traceable to those sources, avoid original research, and write for a general reader. If a topic is contentious - or if your edits could be viewed as self-serving - raise concerns on the article's discussion page. Disclose conflicts of interest and don't edit as if you were an outside observer.
The goal isn't to turn an article into a summary of your own work. It's to make the public-facing account more accurate and more useful.
2. Teach students to contribute, not just consume
A course assignment can ask students to improve a Wikipedia article, assess its sourcing, or propose changes through the article's discussion page. When it's done well, this turns information literacy into something concrete. Students have to distinguish primary from secondary sources, check whether claims are supported, and follow copyright and attribution rules.
The public audience changes the work, too. Students aren't only writing for an instructor who already knows the subject. They have to make the explanation clear to readers who aren't in the course. That can strengthen motivation and reveal gaps in students' own understanding.
Wiki Education, a nonprofit that supports classroom assignments in the United States and Canada, reports that more than 152,000 students across over 8,000 colleges and universities have contributed more than 125 million words to Wikipedia since 2010. Those numbers suggest this kind of educational contribution can be organized at scale. Still, not every assignment will go smoothly: students need training, a manageable scope, feedback, and enough time to learn the editing environment.
Wikipedia's current rules also restrict AI-generated writing. Instructors should check the relevant policy and the assignment requirements rather than assuming generative AI can draft material students can paste into articles. If expectations are clear and students are taught to use sources responsibly, that constraint can make the assignment better for practising independent research and writing.
3. Cite Wikipedia transparently when it was a source you used
Some academics treat any citation to Wikipedia as evidence of weak research. That's too blunt. Wikipedia isn't a replacement for primary research or specialist literature, and it shouldn't be treated as the final authority for a consequential claim when stronger sources are available.
But if Wikipedia materially informed your work - say, by helping you find a term, locate a reference, or understand how a topic is presented to the public - transparent citation can be more honest than hiding that step. Follow your discipline's standards, cite original sources when you rely on them directly, and avoid a blanket "cite Wikipedia everywhere" habit. The point is to make your research process legible and avoid a double standard where AI tools are disclosed but other information sources aren't.
4. Share practical guidance with colleagues
Many scholars want to help, but don't know where to start. A short workshop, an editor-in-residence session, or a collaboration with a campus library can lower the barrier. Keep the first step practical: how to evaluate an article, add a citation, use a discussion page, and handle potential conflicts of interest.
In many real-world projects, the most effective contribution is the one people can repeat. A one-off event may create interest, but a supported course, a recurring workshop, or a library partnership gives participants a path to keep going.
Why universities need to make room for public knowledge work
Individual effort won't be enough if institutional incentives keep pushing in the opposite direction. Academics are usually evaluated on research, teaching, and departmental service. Public contributions like Wikipedia editing often get little formal recognition, even when they improve access to reliable information and support meaningful student learning.
This is, in part, an institutional design problem. Universities shouldn't ask academics to contribute to public knowledge while treating that work as invisible. If a department wants this activity to continue, it should make expectations clear, offer practical support, and avoid assuming unpaid extra labour as the default.
What institutional support can look like
Universities don't need to start a large new program before taking action. They can begin with modest, operational steps:
- Recognize well-documented public scholarship in workload and promotion guidance.
- Offer training through libraries, teaching centers, or subject departments.
- Provide course design support for Wikipedia-based assignments.
- Allocate time for public knowledge contributions when they fall within an academic's role.
- Set clear guidance on conflicts of interest, copyright, sourcing, and AI use.
- Assess impact based on quality and educational value, not just edit counts.
These measures also help prevent a common failure mode: importing institutional authority without respecting community norms. Universities should support academics in learning how Wikipedia works, rather than assuming expertise or status automatically grants the right to steer it.
There's also a larger institutional stake. Public trust in universities depends in part on whether people can see academic knowledge serving the public, not only circulating inside academic publishing. Supporting open knowledge won't fix every source of mistrust, and it shouldn't be treated as a communications exercise. It's a practical way to connect research and teaching to an information resource used well beyond campus.
Conclusion: support the system, not just the page
Wikipedia's long-term health depends on more than whether a single article happens to be correct today. It depends on the people, rules, and resources that make correction and maintenance possible. Generative AI makes that system feel more visible - partly because it can benefit from public knowledge while also nudging users toward answers that may not lead them back to sources.
Academics can respond with focused editing, classroom projects, transparent research practices, and advocacy for institutional support. Universities can make these contributions viable instead of treating them as invisible service. AI companies also have a role here: access to knowledge shouldn't be treated as the same thing as a right to weaken the communities that maintain it.
A practical test is simple: pick one contribution you can sustain - improve a sourced passage, design a supported student assignment, or ask your institution how public scholarship is recognized. Start small, follow community rules, and make the work repeatable. Open knowledge lasts when people invest in the system behind it.