Women need to be put in their place. Women cannot be trusted. Women shouldn’t have rights. Women should be in the kitchen. …
You might have come across the latest UN Women awareness campaign. Originally in print, it has been spreading online for almost two days. It shows four women, each “silenced” with a screenshot from a particular Google search and its respective suggested autocompletions.
Researching interaction with Google’s algorithms for my phd, I cannot help but add my two cents and further reading suggestions in the links …
Guess what was the most common reaction of people?
They headed over to Google in order to check the “veracity” of the screenshots, and test the suggested autocompletions for a search for “Women should …” and other expressions. I have seen this done all around me, on sociology blogs as well as by people I know.
In terms of an awareness campaign, this is a great success.
And more awareness is a good thing. As the video autofill: a gender study concludes “The first step to solving a problem is recognizing there is one.” However, people’s reactions have reminded me, once again, how little the autocompletion function has been problematized, in general, before the UN Women campaign. Which, then, makes me realize how much of the knowledge related to web search engine research I have acquired these last months I already take for granted… but I disgress.
This awareness campaign has been very successful in making people more aware of
the sexism in our world Google’s autocomplete function.
Google’s autocompletion algorithms
At DH2013, the annual Digital Humanities conference, I presented a paper I co-authored with Frederic Kaplan about an ongoing research of the DHLab about Google autocompletion algorithms. In this paper, we explained why autocompletions are “linguistic prosthesis”: they mediate between our thoughts and how we express these thought in (written) language. So do related searches, or the suggestion “Did you mean … ?” But of all the mediations by algorithms, the mediation by autocompletion algorithms acts in a particularly powerful way because it doesn’t correct us afterwards. It intervenes before we have completed formulating our thoughts in writing. Before we hit ENTER.
Thus, the appearance of an autocompletion suggestion during the search process might make people decide to search for this suggestion although they didn’t have the intention to. A recent paper by Baker and Potts (2013) consequently questions “the extent to which such algorithms inadvertently help to perpetuate negative stereotypes“:
It is not possible to know how many people have typed in stereotyping questions about various social groups, and we do not know if such people represent the majority in a population. As noted above, we would guess that actual numbers of people asking such questions are relatively low, but those who do, tend to ask the stereotyping ones. However, even if it emerges that many people are interested in such questions and click on the auto-suggestions that appear, is there an over-riding moral imperative to remove these auto-suggestions?
Google’s autocompletion has been around for quite some time (almost 9 years to be exact, although the official roll-out was in 2008 only). According to the company, the function suggests what it deems “useful queries” (without defining “useful”, bien sûr) to users “by analyzing a variety of characteristics of your custom search engine”. The volume of searches for a specific search term (from different locations) seems to be the main determinate. But there is no reason to believe that suggestions aren’t, to some degree, personalized (the way search results are). And: autocompletion isn’t entirely automated. Google influences (“censors”, some say) autocompletion globally and locally through hardcoding, be it for commercial, legal or puritarian reasons. (Bing does so, too.)
There is no “veracity” to be established because Google is not the objective mirror it claims to be. Google is a company that works however it chooses. Instead of veracity, let’s focus on accountability. Because Google, the very existence of Google, as well as the specific way it works, is having an impact on our lives.
Who is in charge when algorithms are in charge?
I am not implying the negative stereotyped search term suggestions about women are Google’s intent – I rather suspect a coordinated bunch of MRAs are to be blamed for the volume of said search terms – but that doesn’t mean Google is completely innocent. The question of accountability goes beyond a binary option of intentionality or complete innocence.
Unsurprisingly, Google doesn’t take any responsiblity. It puts the blame on its own algorithms … as if the algorithms were beyond the company’s control.
The Spiegel wrote (about another autocompletion affair):
The company maintains that the search engine only shows what exists. It’s not its fault, argues Google, if someone doesn’t like the computed results. […]
Google increasingly influences how we perceive the world. […] Contrary to what the Google spokesman suggests, the displayed search terms are by no means solely based on objective calculations. And even if that were the case, just because the search engine means no harm, it doesn’t mean that it does no harm.
If we, as a society, do not want negative stereotypes (be they sexist, racist, ablist or otherwise discriminatory) to prevail in Google’s autocompletion, where can we locate accountability? With the people who first asked stereotyping questions? With the people who asked next? Or with the people who accepted Google’s suggestion to search for the stereotyping questions instead of searching what they originally intended? What about Google itself? …
Of course, algorithms imply automation. And digital literacy helps understanding the process of automatation – I have been saying this before – but Algorithms are more than a technological issue: they involve not only automated data analysis, but also decision-making (cf. “Governing Algorithms: A Provocation Piece” #21. No, actually you should not only read #21 but the whole, very thoughtprovoking provokation piece!). Which makes it impossible to ignore the question whether algorithms can be accountable.
In a recent Atlantic article, advocating reverse engineering, N. Diakopoulos asserts:
[…] given the growing power that algorithms wield in society it’s vital to continue to develop, codify, and teach more formalized methods of algorithmic accountability.
Which I think would be a great thing because, at the very least, this will raise awareness. (I don’t agree that “algorithmic accountability” can be assigned à priori, though). But when algorithms are not accountable, then who is? The people/organization/company creating them? The people/organization/company deploying them? Or the people/organization/company using them? This brings us back to the conclusion that the question of accountability goes beyond a binary option of intentionality or complete innocence… which makes the whole thing an extremely complex issue.
Who is in charge when algorithms are in charge?
Oh, and of course algorithms are not simply “bad”. The proof: Google Autocomplete can also produce nice things, e.g. poetry.