We’re rapidly moving into an “algorithmic world”and that appears to be beyond dispute. Articles appearing in scholarly journals and popular publications over the last few years make it clear that our interactions with businesses and each other are mediated by algorithms that decide everything from what products we might need to who we might want to be friends with.
In many cases, the use of algorithms provide a great benefit to the user–imagine, for instance, how much more difficult a Google search might be without the “did you mean?” functionality to correct our spelling mistakes and suggest alternatives to our choice of keywords. Or, if we had to search through all the Netflix movie offerings by title or category without a “Because you watched…” feature to make suggestions based on our viewing history.
The downsides of an online world in which algorithms guide the data we are presented and the choices we make have also been addressed at length. In response, some businesses have returned to (or added) human curation of content as an adjunct to algorithms.One downside that is still being addressed is that of “algorithmic transparency”–defined as openness about the purposes, structure and underlying actions of the algorithms used to search for, process and deliver data. — Whatis.com
In the United States, the Federal Trade Commission (FTC) created a new office, the Office of Technology Research and Investigation, among whose charter issues is that of understanding the use of algorithms in consumer-facing applications and how to best avoid harmful consequences while protecting business interesting in protecting intellectual property. Similarly, UK and French groups, among others, are examining this issue in a European context.
One aspect of algorithmic transparency, in my opinion, has not been adequately addressed–that of access to an understanding of the internal workings of these algorithms and the legal and ethical constraints on such access. The issue here arises in the intersection between two trends: the move to proprietary algorithms as the mediation mechanism for much of our online interactions, and the existence of legal constraints to the “probing” of those algorithms by third-parties.
Proprietary algorithms rule our world…
But will we be allowed to understand those?