European Union member-state governments are increasingly aware of the danger of terrorism perpetrated within their own borders - sometimes by their own citizens. From late 2005 onwards, the European commission and justice and home affairs (JHA) council of ministers have rightly begun to place a high priority on curbing radicalisation and recruitment into terrorism, particularly on the internet. The latest manifestation of these efforts is the presentation by the European commission on 6 November 2007 of a new "counter-terrorism package". Its proposals will be voted upon at the next JHA council on 8-9 November.

Johnny Ryan is senior researcher at the Institute of European Affairs (www.iiea.com), a policy think-tank in Dublin with offices in Brussels. He is the author of Countering Militant Islamist Radicalisation on the Internet: A User Driven Strategy to Recover the Web (Institute of European Affairs, 2007).

His blog is here Also by Johnny Ryan in openDemocracy: "The militant Islamist call and its echo" (1 August 2007)The target is evident. Prospective militants can make use of the internet to establish various forms of online, private, person-to-person or group communication - including chat-rooms, blogs, websites, and forums as an alternative (or preliminary) to travelling to Pakistan or other theatres of jihad to gain fighting experience. The question of how to deal with violent radicalisation and recruitment of prospective terrorists within the EU has been a priority issue at successive JHA council meetings since the London bombings of July 2005.

However, the options available to European policy-makers are severely limited. One possible approach has been on the agenda since August 2006, following the disruption in Britain of a plot to destroy transatlantic airliners in flight with on-board liquid-bombs. This involves censorship of the internet. It would be, in my view, an error of major economic and legal significance.

A European doctrine

Franco Frattini, European commission vice-president with responsibility for security and justice, has twice (in August 2006 and September 2007) announced this intention to propose a number of measures, which amount to internet censorship. On neither occasion has Frattini's initiative come before the JHA council for a vote; but the trend of developments in the EU discussion on violent radicalisation is in his direction. The new round of proposals, due to be voted on at the imminent 8-9 November 2007 council meeting, indicates that the commission may be continuing to edge towards unwarranted and impracticable restriction of cyberspace.

The initiative is part of a wider package of anti-terrorism proposals that include a "framework directive" intended to bring into European Union law the Council of Europe's convention on preventing terrorism (2005). Articles 5, 6 and 7 of the convention (which came into force on 1 June 2007) provide for criminalisation of a number of acts:

* "public provocation to commit a terrorist offence" (which the convention defines as "making available ... a message to the public, with the intent to incite the commission of a terrorist offence...")

* "recruitment for terrorism"

* "training for terrorism", including "provid[ing] instructions in the making or use of explosives, firearms or other weapons ... or specific methods or techniques, for the purpose of carrying out or contributing to the commission of a terrorist offence...".

Article 8 notes "for an act to constitute an offense as set forth in Articles 5 to 7 of this Convention, it shall not be necessary that a terrorist offence be actually committed".

While internet-service providers (ISPs) will not be liable for criminal material that transits their networks (as they were not liable under the e-commerce directive), this latest proposal appears to be a continuation of the trend towards repression of free communication across the internet; a trend that is likely to gather momentum if and when a European member-state next suffers a high-profile terrorist attack.

A triple strategy

If Franco Frattini does make a proposal on the issue, it will refer to one or a combination of three forms of internet censorship that he has previously referred to:

1. removing websites from the internet

The removal of (especially) militant Islamist websites from the internet might appeal as an obvious response to a real problem. In July 2006 the EU counter-terrorism action plan was updated to include "acting in common against extremist websites". However, this "solution" is essentially superfluous. The global nature of the internet makes common EU action to take down websites both technically impossible and legally irrelevant in the absence of a binding international treaty and an attendant consensus on what material should be subject to removal. At present there is nothing to prevent websites prohibited within the EU from migrating to a hosting service elsewhere in the world, thereby continuing their presence on the internet.

2. cooperating with ISPs to filter internet access

EU member-states, in agreement with internet-service providers, could attempt to prevent EU internet users gaining access to radical websites using a method known as URL filtering. While this is technically possible, as demonstrated by the "hybrid URL" filtering systems for child pornography in Britain, Norway, and Finland (among others), filtering can be easily circumvented by determined internet users. In addition, the technology required is expensive and can be expected to become more so as the amount of traffic grows exponentially in coming years. EU member-states would face a considerable political hurdle in drawing up common lists of websites to be censored.

Censorship judgments on more complicated content such as web-forum conversations, if new technology were at some point to enable this, would be a particularly difficult legal proposition. Nor would this solution censor individual posts and messages on the chat-rooms and web-forums where militant Islamist sympathisers congregate on the net. Although this gap in censorship could be solved by deploying a combination of "dynamic" or artificial-intelligence filtering and human oversight on all chat-rooms and web-forums operated within the EU, this solution would be very expensive indeed and would make many forums financially untenable.

3. cooperating with internet search-engines to filter results

A further effort could be to oblige internet search-engines to prevent certain search phrases from producing results. In France, Google's search-filters censor Nazi content from among its results. This appears to be practicable but porous, and does not prevent internet users from accessing the restricted material once they find it by other means.

A policy blowback

The pursuit of one or other of these schemes might induce an illusion of security, while in fact it introduces new vulnerabilities to the internet. It is likely that a censorship system would (through human or technical error) block some other material, creating "false positives" whose results include expensive litigation, injury to free speech, and degrading the value of the internet to users in the European Union.

Filtering systems, while expensive to operate and maintain, could also have significant costs for the economy by introducing complexity into what is currently a straightforward and reliable information system. Hybrid URL filtering in particular would impose significant costs on ISPs. Perhaps more important, filtering technology could alter internet users' behaviour and hamper user-driven innovation. Users may fear that attempting to visit prohibited sites will arouse the suspicion of authorities. Or conversely, the prohibition of content could lend it a veneer of rebellious glamour, attracting to restricted material readers who would not otherwise be interested.

The question of censorship raises three further important questions:

* would the loss of signals intelligence to the security services - who would otherwise monitor suspected militants if they were allowed to continue their activities online - be offset by the benefits of censoring harmful material?

* would censorship of the internet be desirable from the perspective of freedom of expression and the unregulated internet?

* Could all of the material that appears to contribute to radicalisation be legally banned?

This last question is particularly relevant since much of the material that contributes to the militant narrative of Islam's persecution by the west is entirely legal and would not be criminalised under the Council of Europe convention referred to above.

The points raised in this brief survey suggest that at the very least, ministers of EU member-states should think very carefully before giving their assent to proposals that may damage freedom while not advancing security: the worst of both worlds.