Detecting Wrongthink Early

1984 taking place in London was very appropriate. The United Kingdom (UK) has become the granddaddy of the surveillance state. Surveilling an entire nation isn’t easy, which is why the UK, like every other surveillance state, is desperately searching for new way to automate its activities. I’m sure that desperation is what lead to this idiocy:

London, United Kingdom – Schoolchildren in the UK who search for words such as “caliphate” and the names of Muslim political activists on classroom computers risk being flagged as potential supporters of terrorism by monitoring software being marketed to teachers to help them spot students at risk of radicalisation.

The “radicalisation keywords” library has been developed by the software company Impero as an add-on to its existing Education Pro digital classroom management tool to help schools comply with new duties requiring them to monitor children for “extremism”, as part of the government’s Prevent counterterrorism strategy.

[…]

The keywords list, which was developed in collaboration with the Quilliam Foundation, a counter-extremism organisation that is closely aligned with the government, consists of more than 1,000 trigger terms including “apostate”, “jihadi” and “Islamism”, and accompanying definitions.

I’m not sure if schools in the UK have deteriorated as far as the schools here but if they haven’t then it’s quite plausible that many of the keywords being looked for would appear quite frequently in a history class. What’s more interesting is that they keywords don’t seem to so much be targeting terrorism as Islam.

It must be noted that using keywords to detect wrongthink is a fruitless endeavor. Because terrorism is currently the biggest target of the State’s propaganda it is a topic of general interest. A lot of people searching for keywords related to terrorism aren’t interested in becoming terrorists but merely want to learn about events related to terrorism. The number of false positives such a system will throw out are going to be far greater than any potentially useful information. Drowning out the signal in noise is counterproductive but it seems to be the strategy most automated surveillance systems rely on.