Child Abuse Images and Cleanfeeds: Assessing Internet Blocking Systems

One of the most important trends in internet governance in recent years has been the growth of internet blocking as a policy tool, to the point where it is increasingly becoming a global norm. This is most obvious in states such as China where blocking is used to suppress political speech; however, in the last decade blocking has also become more common in democracies, usually as part of attempts to limit the availability of child abuse images. Numerous governments have therefore settled on blocking as their 'primary solution' towards preventing such images from being distributed (Villeneuve 2010). Child abuse image blocking has, however, been extremely controversial within the academic, civil liberties and technical communities, and this debate has recently taken on a wider public dimension. At the time of writing, for example, public pressure has forced the German Federal Government to abandon legislation which would have introduced a police run system while the European Parliament has also rejected Commission proposals for mandatory blocking (Baker 2011; Zuvela 2011). Why have these systems been so controversial? Two lines of criticism can be identified, which might be termed the practical and the principled. The practical argument claims that blocking is ineffective, with ill-defined goals and easily evaded by widely available circumvention technologies (see e.g. Callanan et al. 2009). The principled argument, on the other hand, is that blocking systems undermine the norms associated with freedom of expression in democratic societies (Brown 2008). This latter argument stems from the fact that blocking sits at the intersection of three different regulatory trends – the use of technological solutions ('code as law'), a focus on intermediaries and the use of self-regulation in preference to legislation – which individually and all the more so collectively create a risk of invisible and unaccountable 'censorship by proxy' (Kreimer 2006; McIntyre & Scott 2008). This chapter introduces and evaluates these claims by examining three prominent examples of child abuse image blocking – the United Kingdom Internet Watch Foundation ('IWF') Child Abuse Image Content ('CAIC') list, the European Union sponsored CIRCAMP system and United States hash value systems. It discusses the operation of each system and the extent to which the critics' concerns are borne out. It concludes by considering the lessons which might be learned for proposals to extend blocking to other types of content.