NY Attorney General forces ISPs to filter Internet

In another bad day for the end to end principle, the New York Times reports that the Attorney General of New York has succeeded in forcing ISPs to filter their users' internet connections. The expressed motivation is to prevent users from accessing child pornography, though this will be trivially easy to circumvent. There are many problems with internet filtering, and I've written a short summary of them (in a different context) for the Digital Rights Ireland blog. But the New York scenario raises one particular problem - whether this form of censorship, implemented and administered by private actors (who will face an incentive to overblock), can be reconciled with the rule of law. The issues raised are very similar to those presented by the UK Cleanfeed system, about which Colin Scott and myself had this to say at the inaugural TELOS Conference last year:

This presents a number of challenges for the rule of law. Even if an individual ISP’s actions can be described as voluntary, the effect is to subject users without their consent to a state mandated regime of internet filtering of which they may be unaware. The Internet Watch Foundation (IWF), which determines which URLs should be blocked, has a curious legal status, being a charitable incorporated body, funded by the EU and the internet industry, but working closely with the Home Office, the Ministry of Justice, the Association of Chief Police Officers and the Crown Prosecution Service. There is no provision for site owners to be notified that their sites have been blocked. While there is an internal system of appeal against the designation of a URL to be blocked, that mechanism does not provide for any appeal to a court – instead, the IWF will make a final determination on the legality of material in consultation with a specialist unit of the Metropolitan Police.

Consequently the effect of the UK policy is to put in place a system of censorship of internet content, without any legislative underpinning, which would appear (by virtue of the private nature of the actors) to be effectively insulated from judicial review. Though the take-up of the regime may be attributable to the steering actions of government, the way in which the regime is implemented and administered complies neither with the process or transparency expectations which would attach to legal instruments.

There is also cause for concern about the incentives which delegating filtering to intermediaries might create. From the point of view of the regulator, requiring intermediaries to filter may allow them to externalise the costs associated with monitoring and blocking, perhaps resulting in undesirably high levels of censorship. But perhaps more worrying are the incentives which filtering creates for intermediaries. Kreimer has argued that by targeting online intermediaries regulators can recruit “proxy censors”, whose “dominant incentive is to protect themselves from sanctions, rather than to protect the target from censorship”. As a result, there may be little incentive for intermediaries to engage in the costly tasks of distinguishing protected speech from illegal speech, or to carefully tailor their filtering to avoid collateral damage to unrelated content. Kreimer cites the US litigation in Centre for Democracy & Technology v. Pappert to illustrate this point. In that case more than 1,190,000 innocent web sites were blocked by ISPs even though they had been required to block fewer than 400 child pornography web sites.

Orin Kerr has more.

Edit (13.06.08): Richard Clayton indicates that the New York Times coverage may be inaccurate. He suggests that what the ISPs have agreed to is limited to removing certain newsgroups and taking down sites which they host - but does not include filtering of sites hosted elsewhere. There's also some confusion as to just what the effect on usenet will be, with Declan McCullagh reporting that in the case of Verizon all the newsgroups in the alt.* hierarchy will no longer be offered.