Lifestyle
Wikipedia's Operator Tells Courts: AI Training Can Be Fair Use, but Chatbots Repeating Articles Should Need a License
According to The Verge, the Wikimedia Foundation, which runs Wikipedia, has filed a court brief taking neither side in the lawsuits by publishers including The New York Times against OpenAI and Microsoft. It argues for treating AI training and AI output differently, and warns that an oversimplified ruling could cost Wikipedia editors useful AI tools or endanger the open web.
About 5 min read
What happened
According to The Verge, the Wikimedia Foundation, the nonprofit behind Wikipedia, filed an amicus brief (an opinion submitted to a court by a party outside the case) on Friday. The brief concerns a set of lawsuits by publishers, including The New York Times, against OpenAI and Microsoft, and it sides with neither party.
According to The Verge, Wikimedia said in the filing that it worries both sides are pushing an overly broad and simplified application of fair use. Fair use is a principle in copyright law that allows others' works to be used without authorization in certain circumstances. Wikimedia fears that if the publishers win outright, copyright holders could "gatekeep" even innovative and transformative uses, meaning uses that turn a work into something new. If the AI companies win outright, it says, that could amount to a green light for companies to ignore licensing terms.
Wikimedia's middle-ground position
According to The Verge, Wikimedia argues that scraping websites (automatically collecting web page data with software) should not be treated as infringement. It also argues that using copyrighted works to pretrain large language models, or LLMs (the AI systems behind services like ChatGPT), should be fair use, even without a license, if they provide "something new and different from the original sources."
But according to The Verge, Wikimedia believes a license should be required when a model pulls copies of content in response to a user's query and delivers them to the user, replacing the reader's engagement with the original. In that case, the AI output merely substitutes for the original and does not transform it enough to count as fair use.
| Aspect | Wikimedia's position (per The Verge) |
|---|---|
| Scraping websites | Should not be treated as infringement |
| Pretraining without a license | Should be fair use when it provides something new and different |
| Pulling and delivering content in response to queries | Should require a license if it replaces readers' engagement with the original |
| What it means for Wikipedia | The original source must be credited, and adapted content must stay under the same Creative Commons license |
Why this matters especially to Wikipedia
According to The Verge, Wikimedia said in the filing that its sites have become one of the most critical sources of data for all AI and machine-learning training, but that these systems also divert visitors and undermine the communities that maintain the sites. Wikimedia also argues that services like ChatGPT reproduce text freely licensed under Creative Commons (an open content license, which Wikipedia's text uses) without honoring the license terms.
At the same time, according to The Verge, Wikimedia noted that large language models are also useful to Wikipedia, helping editors fight vandalism, detect copyright violations and the publication of others' personal information (doxxing), and find articles that need improvement.
The parties' positions
According to The Verge, The New York Times argues for a more expansive definition of fair use as it applies to large language models and alleges that OpenAI infringed its copyright by scraping its website to build training datasets. OpenAI and Microsoft both argue that using large language models is fair use; OpenAI has said that regurgitating work "is a rare bug."
According to The Verge, Microsoft said in a September filing that expert analysis found Copilot rarely reproduces full sentences from news content used to ground the model, that is, content the model draws on for its answers. Microsoft said this "hardly undermines the transformative purpose of LLM training."
Wikimedia's statements
According to The Verge, Stephen LaPorte, general counsel of the Wikimedia Foundation, said: "A ruling that does not apply fair use principles carefully could threaten the sustainability of free knowledge, and stifle the development of open source technologies, ultimately impacting access to information for everyone in both cases." Wikimedia CEO Bernadette Meehan said that expanding access to knowledge through AI is a powerful goal, but it must respect the people, contributors and communities that make reliable knowledge possible.
In short, Wikimedia wants courts to treat AI training and AI output separately. The Verge's report mentions only the brief and the parties' claims, and does not say that a court has issued a ruling.
Frequently asked questions
What is an amicus brief?
It is an opinion submitted to a court by a party outside the case. According to The Verge, Wikimedia sided with neither party in these lawsuits and instead set out its view of fair use.
Does Wikimedia support OpenAI or The New York Times?
According to The Verge, neither. It believes both sides oversimplify fair use, and argues for drawing a line between training and output.
Has a court issued a ruling?
The Verge's report mentions only Wikimedia's brief and the parties' claims, and does not say that a court has issued a ruling.
Why is Wikimedia worried?
According to The Verge, Wikimedia says AI systems divert visitors and undermine the communities that maintain its sites, and that some services reproduce Creative Commons-licensed text without honoring the license terms. It also worries that an overly black-and-white ruling could cost editors useful AI tools.
Browse the latest news in this topic
Lifestyle
AI Data Center Backlash in Maine Becomes a Senate Race Flashpoint, The Verge Reports
The Verge reports that many Maine residents worry AI data centers raise power bills, drain resources and bring few lasting jobs. That backlash has become a Senate race issue, and the result will help decide which party controls the Senate and shape US data center and AI policy.
Lifestyle
ElevenLabs Launches Eleven Realtime: Building Voice AI Without Assembling Every Component Yourself
ElevenLabs has announced Eleven Realtime v1 Mini and v1 Max, saying developers can build AI that talks by voice with a single model, without choosing and tuning many components. This may cut work for teams building voice support or assistants, but the API supports WebSocket connections only.
Lifestyle
No Belief in a God Required? A Humanist Wins an Exemption From Using AI at Work
According to TechCrunch, an administrator at the University of Kentucky has obtained a religious exemption from using AI at work, based on humanist beliefs. This piece covers what happened, the university's position and what it means for ordinary workers.
Lifestyle
Amazon CEO: AWS Could Become a $1 Trillion Business, but First Comes About $220 Billion in Spending This Year
According to TechCrunch, Amazon CEO Andy Jassy said on the earnings call that AWS, Amazon's cloud business, could reach $1 trillion in annual revenue within five to ten years. Heavy AI infrastructure investment comes first: cash capital expenditures this year are expected at about $220 billion, and investors are concerned about the spending.