In many ways, for many purposes, the internet provides the most accurate resources of information ever available to human beings. A case in point is Wikipedia. Let us first admit that many valid criticisms can be made of Wikipedia. On unpopular or relatively obscure topics, Wikipedia merely reproduces seriously dated publications that are in the public domain. On some topics, the presentation is shaped by amateurs with uninformed and peculiar points of view. Topics can be hijacked by political operatives. Information from more authoritative sources is not always accorded greater value. There is no systematic practice of fact-checking. Many of the entries read as if they were composed by disorganized committees of volunteers (which they are). And more criticisms can also be made; a long list of them, in fact, can be found in the Wikipedia entry entitled “Criticisms of Wikipedia.”
But despite these criticisms, and despite persistent injunctions of college professors against using Wikipedia as a source, it is without a doubt the greatest single source of knowledge ever assembled. The more responsible epistemic agent will always balance whatever is said on Wikipedia with a broader survey of other more authoritative sources, but for a quick and mostly accurate overview of the widest array of possible topics, no other encyclopedia even comes close. Even if Wikipedia is a second-rate (and sometimes third-rate) resource on each particular topic, there is no other base of knowledge that can come close to its range of coverage and general level of accuracy.
And, of course, Wikipedia is not the only available source of knowledge on the internet. The internet gives us access to first-rate scholarly journals, news media, blogs by true experts of obscure matters, maps, lectures, and so on without limit. The internet makes it easier than ever before to take a broad sampling of different accounts of nearly any topic, and to form judicious opinions based upon that diversity of resources.
But this is only, at best, half of the story. As recognized at the end of the previous section, internet searches can be “gamed,” and there are powerful incentives for distorting what users find as a result of their searches. This is shown most dramatically in instances of exploiting data voids.
As expansive and comprehensive as the internet is, there are topics that have a very minimal presence on the web. It may be a set of words that is not commonly used, or it may be a person or event or little town about which people have very little to say. We can call such a neglected entity a “data void,” which simply means that there is not much information on that particular topic on the internet. If, for whatever reason, someone wishes to tell some particular story about these data void entities, they can tell that story in multiple places throughout the web, and mindless algorithms will direct users to that story if they happen to search for that entity. A data void is thus an opportunity to establish and control a narrative.
For example, the term “crisis actor” was for some time a data void. No one searched for that pair of words. But, according to recent media and internet researchers, at some point malicious individuals seized the term and populated the internet with many false stories about people who were hired to pretend to be victims of mass shootings. Multiple websites offered seemingly genuine accounts by “crisis actors” who admitted to having portrayed victims of various faked or staged mass shootings. Efforts were then made by the malicious individuals to get the phrase “crisis actor” mentioned on some national media outlet. The efforts succeeded, and when viewers of the media went to search the term “crisis actor” they found multiple accounts from different sources of people admitting to playing roles in staging fake shootings. This gave support to baseless conspiracy theories about the government staging mass shootings. Similar data void hijackings have been executed with the terms “collusion hoax,” “black on white crime,” and “pizzagate.”
This is an example of malicious agents exploiting the mindless operations of algorithms to bring baseless conspiracy theories to a broad audience’s attention. But there are also less outrageous attempts at doing the same thing. Political organizations can promote specific terms and slogans and make sure that they direct the public narrative by establishing websites that become the go-to sites for searches employing those terms: this is called strategic keyword signalling. These are efforts to “game” the mindless functioning of algorithms so as to exert influence over what broad communities take to be truth.
Of course, as a student in epistemology that has thought through skepticism and the Grand Deception Doubt, you will naturally wonder whether “strategic keyword signalling” is itself an exploitation of a data void. Perhaps the experts in media studies are doing exactly what they are accusing the people behind “crisis actors” as doing! And a terribly destructive seed of skepticism is thereby planted. We may begin to suspect that we can no longer trust anything we find on the web. “Anyone can ‘prove’ anything by posting false information and manipulating the internet’s algorithms,” someone might think. “We cannot ever know anything, and so we might as well choose to believe whatever story we like best.” These extreme doubts will only be nourished by our observation of how much power and wealth there is in the control of algorithms. There are strong incentives to control the knowledge of individuals, and the private companies hosting search engines have unparalleled power and incentives to exert that control. With so much at stake, how can information not be thoroughly biased and skewed toward the interests of those in power?
But as students of epistemology, we also know how to begin to think our way through these extreme doubts. If we adopt a basic Humean or Bayesian outlook on the information we are coming across, we shall start to assess the occasions where we have solid reasons for doubt and those where the reasons are less solid. The claim that public shootings are entirely fabricated as the result of massive government conspiracies is an extraordinary claim and should require extraordinary evidence—far more extraordinary than a small collection of obscure websites. The claim that everything on the web is fabricated is even more extraordinary, and finding evidence for such a claim through conspiracy-theory websites leans decidedly in the direction of being a self-refuting justification. The more plausible claim is that there is genuine information and disinformation on the web, and patient inquiry and reasoning is required to sort the more likely from the less likely or the reliable from the unreliable—as has always been the case in human knowledge.