Democracy of the Algorithm
Working on the Society of the Query conference, I often find myself confronted with an unquestioned believe in what are believed to be the empowering or even emancipating qualities of universally accessible open and free information. Michael Stevenson once referred to this believe as Information Determinism, as if information by itself will solve social and political issues. Realizing universal availability of information is difficult without being responsive to local condition, but in including national rationales, different views on what is considered harmful information are introduced. A much heard outcry of critics of Google’s censorship practices in China is that one only need to search for keywords such as “Tiananmen Square” in both Google.com and Google.ch, and see in one of them many iconic images of protesters and the crackdown itself and in the other tourist pictures of the square lighted up at night and happy Chinese couples posing before it.
Although claims about China’s violation of human rights seem unquestionable, it is ultimately a political and social issue and not a technological one. Basing arguments about Chinese repressive practices on accessibility to Google search result, or more generally, what is presented in the search results, is problematic and fails to look at what Google is. The nature of Google’s search results is based on popularity vote and thus can be seen as a reflection of the implicit recommendation of the dominant Web users. Although Google’s algorithm has often been heralded as democratic, receiving dominance through number of inlinks (also referred to as votes), not all sources are treated equal; a link, although being a popularity vote, does not necessarily mean an agreement on principles. For most part of the Web’s history the dominant users on the Web were western users, and thus what is reflected in the Google search results might not be a reflection of Chinese users and their opinions. If it was, it might be insightful to think about whether criticism about Tiananmen Square would surface if the search results included all the voices of the Chinese users. Maybe more importantly in relation to the Chinese identity, would the majority of the user like to have such a national drama be represented on the top of the Google results, for the entire world to see? What does it say about the state of local social and political issues when they are not reflected in a global publishing platform such as Google?
These questions are hard to answer, but what is clear is that at this point in time the voices of the Chinese users might not be represented equally in the search results. Partially this is due to the fact the Chinese user, although large in numbers, is still the new kid on the block. Another reason is that not all voices are heard, not all the different views on reality are being allowed a place in the online debate. But assuming for the moment that Google would be able to operate the way they think is best, operating according to their mission statement, would those voices be reflected in the search results? Keeping the analysis close to home, the question could become whether or not the search results we encounter are a reflection all our voices.
Richard Rogers in ‘The Googlization Question, and the Inculpable Engine’, looked at the search results for queries concerning important western political and social issues and found that rather then providing a collision space of alternate account of realty, something you might expect in a democratic society, Google furnished the familiar, it literally returns what you would expect. The search results follow mainstream storylines, issues raised in mainstream media, who have been repeated frequently on television, made up the top of search engine results. For Rogers Google is a status-authoring device and explains that:
Given all the pages that do reference a key word, the search engine delivers those ‘deserving’ to be listed as the top sources. Thus, apart from seeing the source set as the story, one also may view the engine results as telling a second kind of story — the current status of the topic or issue in question through the organizations currently representing it, on the record, in the engine returns.
Seeing the Google search engine as producing information that represents the diverse sets of opinions needed for true democratic debate (if we for the moment assume this is at all possible through media) thus ignores the way the technology operates. It is not so much Google search engine serving a misleading presentation of facts, but more so a misunderstanding of what it is. The search results are returning exactly what could be expected and the service works just fine. In a media saturated environment it should not be surprising that what is reflected in a service based on popularity votes is a reflection of issues and opinions brought up by the dominant media. Given the dominance of mainstream media in the Google results it seems a plausible assumption that a similar situation would occur in China. Given that the mainstream media operate under the same self-censorship regulations as Google.ch has to abide to, a free Google might change less then one might assume. In the act of trying to free the Chinese people from the oppressive governmental censorship, western users might be loosing sight on how their own system is reproducing biased narratives or at least return only a selection of a very particular reality.
In the article Google’s China Problem (and China’s Google Problem), Clive Thompson from the New York Times spoke to a Internet executive highlighting the “distorted universe” the west believe the Chinese people are being presented and wonders:
What happens to people’s worldviews when they do a Google search for Falun Gong and almost exclusively find sites opposed to it, as would happen today on google.cn? Perhaps they would trust Google’s authority and assume there is nothing to be found.
Thompson reflects on this and states:
Perhaps the distorted universe is less of a problem in China, because — as many Chinese citizens told me — the Chinese people long ago learned to read past the distortions of Communist propaganda and media control.
What happens when we read Thompson’s conclusion slightly different? What if the western faith in technology and information, in the belief that the information returned are all the voices, that universal access to al information will change complicated social issues, have made us more blind to the constructedness of our own reality then the Chinese to theirs? In the act of trying to free a country of its oppressive regime by providing them access to information, we assume they are blind to something they have been accustomed to for decades and ignore our own blindness.