Researchers from the University of Cambridge are developing a new capability for drones: identifying violent incidents within a large audience - even before they develop into a fight or attack. The idea sounds quite fictional, but it relies on three well-tested technologies: a high-quality camera, a machine vision software that will recognize each person in the crowd and his hand movement, and an artificial intelligence tool that can recognize when Adam A is pushing Adam B in a way that can lead to a fight . The development will be examined this summer at festivals in India.
The idea is to improve public safety at mass performances and festivals, by flying eyes to watch over the celebrants. Will it work? I bet not, for several reasons. First of all, people are physical beings and it is unclear how the software will differentiate between someone who has given fun to his friend and Mecca. How would you differentiate between a couple who sneezes for pleasure and a case of assault? Between someone who accidentally bumps into another and falls down and deliberately falls prey?
"Come fast, there are fights120 m southwest of the stage"
People are also physical in a different way; For example, it seems that a group of Israelis at one of the Indian festivals (you and I will be there Israelis) will be more physically than a bunch of Norwegians. And for the moment, even if the software detects a violent incident within a large audience - how long will security guards arrive there? How easy would it be to guide them there? ("I see the huge stand of the speakers, around 3,500 people, 50 miles to the west).
No, to be effective, the skimmer should also include a face recognition algorithm. Which will attach bio-metric images to the database. Otherwise, the authorities will not be able to know who the villain slapped a poor man who accidentally dumped a beer at Iron Maiden. And here we are already undergoing development that is designed to improve public safety to the one designed to undermine and frighten it. I said this once, I say it again: Anyone who wants a safe society should educate against violence and not intimidate the public. It takes longer than improvising technological solutions, but pays off in the long run.