12
votes
In response to mass shootings, some schools and hospitals are installing microphones equipped with algorithms
Link information
This data is scraped automatically and may be incorrect.
- Title
- The Unproven, Invasive Surveillance Technology Schools Are Using to Monitor Students
- Authors
- Jack Gillum
- Published
- Jun 25 2019
- Word count
- 1964 words
Yet another rush to commercialize an unproven machine intelligence product as a surveillance technology, with predictably poor results. The question remains, if it was effective, would we want it, and shouldn't those surveiled be informed?
Personally, I'm not sure of the answer to the question of whether I'd want to live in an environment that monitors for "aggression". At what point does dissent or dissatisfaction make you a target?
In Wim Wenders' 1997 film, The End of Violence, he predicts a space-based surveillance system capable of identifying violent acts; it's close to infallible, but the populace surveiled isn't allowed to know it exists. Shadowy government agents kill or otherwise neutralize those who might disclose the system's existence. [I don't recommend this movie as the plot is a jumbled mess and the acting is execrable, but the concept stands.]
Yet another profit-seeking company willing to sell subpar products to schools who are desperate for solutions in a culture of crisis.
The sad reality is that I see how these could be beneficial--less for the aggression detection and more for firearms. Unequivocally identifying gunshots the moment they happen would mean that instantaneous notification and response would be possible, which would undoubtedly save lives. Combine that sort of detection method with something like fire doors that close and lock, and you could contain a shooter to a limited area the moment they initiated.
Aggression detection, as noted in the article, is much more a fool's errand. There are plenty of forms of aggression it won't be able to detect, and the moment these sensors become widespread, people will simply adapt their behavior to not trigger them. That's also assuming they can ever get the detection to a satisfactory level of accuracy. The detection in the videos was so poor that I could have produced similar results by just programming the microphone to note relative peaks in volume and then flip a coin for each.
Even if the microphones can adequately detect aggression in an individual voice, I seriously doubt that they're able to do so with fidelity when listening to a large group. I would wager that even the most sophisticated equipment couldn't do real-time processing of the tumultuous ocean of noise created in the average school cafeteria during lunch.