Hickory dickory Grok, AI on the block
Wikipedia has never been the favorite website of American conservatives, but a new initiative from the Heritage Foundation seeking to "identify and target" editors is a concerning development.
📰 In the News
Grokipedia opens to general public; General public not interested
After years of lambasting Wikipedia, Elon Musk walked the walk and launched Grokipedia last month. But will the encyclopedia fueled by the X/Twitter generative AI chatbot Grok deliver on its promises to fix Wikipedia’s many problems?
That may depend on who you ask. Early returns show Grokipedia content has its own biases that more closely align with the worldview of its founder. Scale is another issue: at the time of publication, it was about one-ninth the size of Wikipedia by number of articles, many of which were copies of its older sibling’s homework.
Grokipedia might be the highest-profile Wikipedia competitor in recent memory, but it isn’t the first. From Conservapedia to Wikigen.ai to Wikipedia co-founder Larry Sanger’s Citizendium, many have tried—but failed—to take on the original, and interest in the newest challenger is already waning.
Traffic to Grokipedia topped out at more than 460,000 visitors on Oct. 28, but after the initial flood of interest, user numbers plummeted. Data from Similarweb showed a near-vertical drop in traffic and Grokipedia is now hovering around 30,000 daily users. For a point of reference, English Wikipedia alone has averaged 230 million daily page views the last few weeks.
Not only are Grokipedia’s views low, but so is the industry’s concern with it. The site is being treated as a novelty rather than a serious source of information.
🔔 Wiki Briefing
Another day, another crawler: Foundation takes on data scraping bots
If you ask the Wikimedia Foundation (WMF), the relationship between its free use data and artificial intelligence companies is looking more parasitic than symbiotic. Now, the Foundation wants to restore the balance.
Wikipedia’s anti-AI stance largely stems from its editors, who time and time again have spoken against generating encyclopedia content with LLMs. The WMF leaves its editorial decisions to the users, and has largely kept quiet on the topic. That is, however, until AI impacts its servers or its wallet. The WMF sounded alarms earlier this year about how information scraping was dealing a blow to its bandwidth. Now, it says it’s time to pay the piper.
The Foundation wants acknowledgement of Wikipedia’s human contribution to these AI models and for companies to pay to access data through the Wikimedia Enterprise platform rather than using bots to troll for information. Time estimated that 65% of the most server-straining traffic comes from bot scraping.
The WMF has made it clear that without humans—especially those working on the project as volunteers—AI can’t exist. Specifically, it can’t exist without the human effort to gather information, which the Foundation says is why “Wikipedia is one of the highest-quality datasets in the world for training AI“.
What’s unclear is how big companies will react to this request, because it is just that: a request. No penalties or consequences have yet been outlined. But if these tech companies aren’t careful, they might just strangle their golden goose.
📚 Research Report
♫ Are they humans or are they bots? ♫
In our last edition, WikiWise reported how London researchers looked for changes in Wikipedia use and readership following the launch of ChatGPT. A correlation was found, showing that even though readership was up, AI search was slowing Wikipedia’s growth in places where ChatGPT was available. The WMF took a new look at their own viewer data, though, and found it may depend on who (or what) is looking at Wikipedia. Specifically, it seems that human page views on Wikipedia are decreasing.
A recent blog post from the WMF reported an estimated 8% decrease in viewers compared to 2024 after changing how it filtered bot traffic. It noted how difficult it can be to distinguish between human and bot views, especially as bots get better at mimicking humans and that this new data “must be treated with care”. Even so, the Foundation attributed the traffic dip to the rising popularity of LLMs and an increased reliance on social video for information.
📖 Jimbo’s Book Club
Wikipedia co-founder Jimmy Wales has been on a media blitz this month, promoting his new book, The Seven Rules of Trust.
But that’s not what his interviewers want to talk about: Wales has faced questions about Grokipedia, criticism of Wikipedia from the American political right, and how the UK’s Online Safety Act might sweep Wikipedia up in its dragnet. The chairman emeritus of the WMF appeared on NPR, PoliticsHome, the BBC, and even before the United Nations, to discuss the book and all of these other issues.
Criticism of Wikipedia isn’t without merit; even User:Jimbo_Wales thinks it still needs work. He’s gone on record saying it needs to more fairly represent the many sides of complicated issues. Wales isn’t shy when it comes to stepping into controversy, either. He called the Gaza genocide Wikipedia entry a “particularly egregious“ failure to neutrally discuss the differing viewpoints on the topic in a manner befitting an encyclopedia.
Co-founder or not, Wales is still just another editor giving his opinion, something the Foundation is sure to remind everyone of, especially at a time when its editorial influence is questioned.
💡 Tips & Tricks
Have a trip coming up? Ditch the travel agent and sail over to Wikivoyage! The free global travel guide is under the Wikimedia umbrella and covers nearly 34,000 travel destinations. From itineraries to language guides to safety advice, this guide that anyone can edit has you covered.
