>>19501Anything more than observation could get OP into trouble. Given what you've said, it sounds like the messenger-nodes have a high signal to noise ratio.
What could be done:
>Observe messenger-node interactions until you can reliably imitate them.
>Decrease the signal to noise ratio by activating imitation messenger-nodes at irregular intervals, or when other nodes, such as T-nodes, interact with them. Your objective could be to slowly increase the number of observable interactions between imitation messenger-nodes and observers which would decrease the visibility of "true" messenger-nodes.
>You could do this in real time, operating large numbers of automated imitation M-nodes. If you had the resources, this could be applied to the point of disrupting the entire network. This could be problematic because you'd be attacking the communication medium itself, which is wasteful and ineffective. The primary objective of imitation is observation. Disruption or misdirection should be considered secondary objectives, and applied only when necessary. Diminishing returns and such.
>When you have a network of imitation M-nodes that can reliably ID T-nodes, you could re-apply the technique to imitate T-nodes as well.
>Over an extended time, one could selectively and slowly imitate the entire structure of "true" M-nodes and T-nodes by replacing them with a significantly larger number of imitation nodes, which do not all need to be actively operated at any given point in time. Given enough time to create and build an imitation recruit network, you could significantly restrict real human supply chains.Careful imitation and strategic misdirection may be more effective than observation alone at disrupting extremist recruitment via social media, particularly when operators have been operating under the assumption that they are under observation. If applied imprecisely in a non-convincing manner it would be a huge waste of resources.
There are drawbacks. Increasing public perception that an extremist threat is "growing" and spreading extremist propaganda across social networks. Both may have unintended consequences, such as provoking a largely disproportionate or inappropriate response to a statistically insignificant but genuine threat. This wastes valuable resources dealing with "counter-extremists" and potentially increases the number of real recruits and sympathizers on both sides.
In this example, the disincentive to non-state actors performing this technique is that they may mislead various state-actors, becoming targets themselves, while "real" extremists may already be operating "in the clear", relatively safe from harm under supportive regimes and operating under the base assumption that they're already targets for observation or enforcement.
Observation is likely to fail as an effective deterrent to those individuals who are working in part to service their own ego or reputation. They may be seeking observation and recognition, which is why imitating them or otherwise stealing the show could be more effective than reactive censorship.
Strict, abundant, reactive censorship, even in small amounts, can amplify the signal to noise ratio of extremist messages or messengers; in systems of mass censorship and surveillance, observers of censored non-extremist content (such as copyrighted or morally objectionable content) can easily become accustomed to circumvention of censorship and surveillance technologies.
Non-extremists driven to unconventional communication techniques in order to preserve their basic human rights to free expression may be more likely to encounter and therefore find something in common with extremists.
