Is Wikipedia Dead?
I think Wikipedia might be on its way out.
For a meta moment à la Inception, check out the Wikipedia Wikipedia page.
Since its founding in 2001, Wikipedia has sought to organise the chaos of information on every topic across the Internet into a single, coherent space.
It hasn't been a straightforward journey. From the page:
"Wikipedia has been praised for enabling the democratization of knowledge, its extensive coverage, unique structure, and culture. Wikipedia has been censored by some national governments, ranging from specific pages to the entire site, sometimes due to its criticism of the government or by content otherwise considered blasphemous. Although Wikipedia's volunteer editors have written extensively on a wide variety of topics, the encyclopedia has also been criticized for systemic bias, such as a gender bias against women and a geographical bias against the Global South."
Even the critics, though, would struggle to claim that the Internet would be a better place if Wikipedia did not exist. It became the go-to source for roughly accurate information on pretty much every concept, both past and present. It has saved many-a-student from bewilderment (including myself), and has been a permanent fixture amongst the top organic results for most knowledge-seeking Google Searches for as long as I can remember.
I think, though, its time might be up.
AI is becoming a living, breathing oracle. No longer are people searching and navigating to a standalone website like Wikipedia to find carefully balanced answers. The AI search result is the answer, whether surfaced within a chat product or as an AI overview on a Google Search query. There is simply no longer a mass-market role for a Wikipedia-like aggregation point.
This is obviously worrisome for a lot of reasons. Wikipedia has multi-party checks and balances on accuracy and balance - an AI answer does not. Wikipedia has an explainable audit trail of content changes - an AI answer does not. Wikipedia has a form of permanence and repeatability, in that two users see the same content for a given concept - an AI answer does not. AI's ills: hallucinations, 'black-box-ness' and sycophancy are all major barriers that need to be overcome to avoid a spiral into a splintered reality of confirmation bias and misinformation.
Maybe I'm wrong. Maybe Wikipedia is such a cherished brand and destination that it will live on as the cornerstone reference point for a 'roughly right' answer on the world. But I fear not.
There is still time, for now, I suspect. These things don't explode in an instant: I'm not predicting a binary light switch flicks us into darkness overnight. But the dimmer switch may start to turn: slowly, at first, but at a quickening pace. Less traffic, less readership, less income, less resource, less scrutiny, less incentive to maintain accuracy, less reliability, and so on. Like network effects and leverage, it works wonderfully on the way up but catastrophically on the way down.
I for one will miss it if/when it goes.