Wikipedia’s Monopoly on “Truth”
-Helen Buyniski
Through a dense web of relationships with Big Tech, legacy and social media, NGOs, educational institutions, and even government agencies, the Wikimedia Foundation has covertly seized the dominant narrative position for the establishment-friendly worldview it propagates through Wikipedia, WikiData and their sister sites. This flies in the face of US anti-trust regulations and threatens the very foundations of consensus reality.
Despite the lip service it pays to the “wisdom of crowds,” Wikipedia as it currently exists is incompatible with the free market of ideas. Politically-sensitive topic areas are heavily restricted if not entirely off limits to ordinary editors, while only a narrow range of viewpoints are even permitted into articles, a spectrum tightly controlled by restricting sourcing to the “safe” legacy media that have always served as a reliable echo chamber for power.
Because Wikipedia is free and shields itself from legal liability for the libels it publishes by hiding behind Section 230 of the Communications Decency Act, its competitors, which have to actually pay their writers and maintain some accountability for what they publish, cannot realistically compete. Even if their product is superior, established encyclopedias like Britannica or World Book face an uphill battle to win the attention of the terminally-online, one that becomes even steeper when Wikipedia’s privileged place in Google search results and AI training is taken into account.
Alphabet (the parent company of Google and YouTube) has so deeply integrated Wikipedia into its offerings that it can be hard to tell where one ends and the other begins. The “knowledge boxes” returned in pre-AI Google searches featured data piped directly from Wikipedia without attribution, giving the user the false impression that the contents were facts distilled from the billions of pages indexed by the search engine. Further cementing the epistemological monopoly, Google uses Wikipedia’s de-facto social credit score to decide how those pages are ranked, heavily weighting the encyclopedia’s opinion of webpages and their authors when they quantify reputation using a metric called EEAT (Experience, Expertise, Authoritativeness, and Trustworthiness). (1)
Google has for years faced credible accusations of operating as a monopoly across multiple sectors, and its alliance with Wikipedia should trouble anyone concerned with freedom of thought. The company’s former CEO Eric Schmidt publicly acknowledged over a decade ago that its ultimate aim was to present the user with a single answer to each search query - one the user has been psychologically manipulated into trusting. Now that Google has shoehorned AI into every search, that dream has been realized. Only 1% of users click through to source websites when an AI summary is displayed atop their search results (2), even though Google’s search AI is notoriously prone to hallucinations, delivering false responses in up to 9% of cases - resulting in a huge number of wrong answers when one considers that Google fields billions of search queries every day.
Like its competitors OpenAI and Anthropic, Google uses Wikipedia to train its AI - indeed, the New York Times calls Wikipedia “probably the most important single source in the training of AI models.” (3) The Wikimedia Foundation actually makes money selling these and other Big Tech firms the content its unpaid editors authored, claiming it represents the last entirely human-generated data-set on the increasingly AI-infiltrated internet even though Wikipedia editors have been using AI to write articles for years. This is no small deception. Training AI on AI-generated content - data-inbreeding - causes a phenomenon known as model collapse, leading to increasingly homogenous, biased, or nonsensical outputs. The unreliability of Google’s AI search - which has already led to a court judgement against the company in Germany, with Google found directly liable for false claims in its AI overviews after it libeled two publishing companies (4) - may be a direct result of engineers training it on Wikipedia data.
Wikipedia doesn’t just supply the raw material for training AI models - it’s their preferred source for generating search results, comprising over 50% of source material in some cases. Its epistemological dominance is so total that internet marketing firms advise their clients that the best way to “reach AI search visibility” is with a Wikipedia article. (5)
As we discussed in earlier segments, Wikipedia only permits mainstream establishment publications as sources for its articles, meaning no competing narratives can enter the information ecosystem by this route. Social media platforms like Facebook and YouTube use it as their fact-checking bible, further marginalizing dissent. Journalists who likely used Wikipedia for research years ago are now using AI to write their articles, ensuring the resulting slop is doubly saturated with the establishment’s preferred narratives.
Even the classroom, once a hallowed Wikipedia-free zone, is no longer safe for heterodoxy. Wikipedia won the allegiance of teachers during the Covid-19 pandemic, infiltrating schools with free educational materials at a time when many faced a sudden need for distance-learning material and could not find (or afford) accredited alternatives. The Wikimedia Foundation’s partnerships with UNESCO, the World Health Organization, and other UN bodies have further legitimized what was once derided as the bathroom wall of the internet. Its alliances with the leading lights of the public and private sector have all but guaranteed Wikipedia a foundational role in the future of the internet itself, a sobering possibility given its flagrant disregard for the truth, which we have exhaustively documented in previous articles.
Once the AI-powered information superstructure becomes widely adopted across all levels of society, it will be all but impossible to redeem one’s reputation if one is denounced on Wikipedia, as there will be no empirical vantage point from which to examine the encyclopedia’s claims. One cannot query a search engine for alternative views if search engines merely ask Large Language Models trained on Wikipedia what Wikipedia says about the topic. The public, initially disincentivized from doing the hard work of seeking out alternatives to the Wikipedia-promoted establishment narrative, will with each successive generation become less and less aware that a reality outside the truncated spectrum of discourse permitted on Wikipedia and its partners ever existed, and humanity will have been effectively entombed in Plato’s Cave.
1 https://nomadicsoftware.com/blog/google-e-a-t-seo-2/
3 https://www.nytimes.com/2023/07/18/magazine/wikipedia-ai-chatgpt.html
5 https://themathergroupllc.com/why-wikipedia-is-the-hidden-lever-for-ai-and-llm-search-results/


Excellent article!
All part of TrustLab, The Trust Project Network and the Trusted News Initiative: "building a safer web for everyone."
https://theylied.ca/MainstreamMedia.shtml
.
Yes - this is absolutely true - Wikipedia is an insidious Big-Pharma propagandist - I had a biographical listing in Wikipedia for many years, until the day in 2015 when I published one sentence in a larger article questioning vaccines, and within a few days I was "disappeared" off of Wikipedia forever.