The internet industry is multi-faceted. In particular there are two main types of participants—Internet Service Providers (ISPs) and hosts.

ISPs provide a conduit that connects consumers to hosts and to each other. Their goal is to deliver an internet connection at the lowest possible cost with highest possible speed. The focus of an ISP is often highly technical, making use of routing and caching technologies to improve customer experience and reduce costs. With regard to radicalisation content, an ISP is often not aware of the content being delivered to subscribers and may not even be able to view it as a result of it being encrypted. An ISP will lose money on customers that: require support frequently, draw data volumes at a level that is near to their ‘cap’, and are in new areas where the ISP has made capital expenditure on equipment that does not yet return enough income to justify the cost. ISPs make money from established customers that operate well within their limits and do not draw on support resources. It is a business of averages and long-term relationships.

Hosts are internet industry participants that provide content for users to consume. Examples are Google and Yahoo. Their income is often derived from advertising. The trend in host service is ‘Web 2.0’ which is a buzz-word for consumer-generated content. With regard to radicalisation, a host may provide a Web 2.0 facility for consumers to start a ‘thread’ and then for others to join and share ideas on the thread.

The regulatory framework in Australia provides ‘safe harbour’ for ISPs but not for hosts. For this reason, much content is hosted off-shore and the majority of hosting contracts prohibit any content that is likely to result in liability or additional costs for hosts. At present the safe harbour legislation is being tested in the case iiNet vs AFACT.

Given the above nature of the industry participants, the internet Industry is relatively risk averse and will avoid radicalisation content because it has the potential to result in liability to the ISP/host and/or result in increased support costs.

A difficulty in engaging ISPs and hosts in the fight against radicalisation is the issue of jurisdictional boundaries. The difficulty arises where the hosted content may be legal in the country in which it is hosted but not legal in Australia. If the content is encrypted then the ISP will have no way of determining what the content is and little ability to determine whether access to it breaks any laws, in particular given that ISPs do not have a mandated regulatory role. It is arguably more valuable and productive for ISPs to work with Australian law enforcement agencies to help track and identify individuals involved in radicalisation rather than attempting to block such content.

The internet industry is generally opposed to ISP-level filtering because:

1. It is not an effective way to block or stop the content that it is designed to prevent.
2. Filters can be bypassed. This would be particularly relevant in radicalisation networks where the participants form a relationship with each other which could involve sharing of techniques to bypass filters.
3. The introduction of (arguably ineffective) filters will increase the costs to ISPs and impede their ability to provide superior performance. This will ultimately lead to higher costs and lower performance for consumers.
4. Legislative change is far too slow to be effective. For example, the time to introduce filtering legislation, or revisions to such legislation is months or years whereas technology to bypass and/or avoid filters would occur in days or even hours.
5. The introduction of filtering legislation would be an impediment to investment and innovation in the Australian internet industry. It is already the case that the lack of safe harbour legislation for hosts within Australia drives Australian content to be hosted offshore. If filtering is introduced this will worsen.

The Australian internet industry is largely supportive of the need for the internet to be used for ‘good’ rather than ‘evil’ and as a result is very prepared to cooperate with law enforcement in any way possible.

Of note is a recent move by Facebook to remove a group called ‘I Hate Muslims in Oz’ because it contained an explicit statement of hate which is a violation of the terms of service of Facebook. This group was created by consumers and only came to the attention of Facebook as a result of user complaints. Ultimately the group did little harm and certainly did not play any significant role in radicalisation but its removal highlights the risk-averse nature of Facebook in both terms of service as well as tolerance for such content.

In forming counter-radicalisation policy it should be remembered that the internet provides a tool to help humans communicate better and more easily. It is a facility that underpins a human activity system. As part of a human activity system, the internet evolves often very rapidly to trends driven by human behaviour. The internet is resistant to impediment—where changes are introduced they are often quickly circumvented.

From an industry perspective, the best way to fight radicalisation is to use human law enforcement, equipped with state-of-the-art technology, assisted by the Australian internet industry.