Have you ever seen the episode of Black Mirror called “Arkangel”? It tells the story of an overly apprehensive mother who implants a chip in her daughter's brain to follow her every move.
But it also updates it with a couple of features, like the ability to see everything your daughter sees and impose automatic visual censorship on her. Basically, he blocks images of anything that might be considered “shocking” to his daughter. Her daughter simply cannot see what her mother censors. Needless to say, this episode of Black Mirror isn't going to go well. It is a prerogative of the English dystopian series, on the other hand.
A technology is being developed today that could actually be used to automatically censor images in real time.
In this clip by TEDx Talks, computer interface scientist Lonni Besançon presents technology that could do just that.
The automatic visual censorship system works slightly differently to the version seen in Black Mirror, aiming to preserve additional information of the image that is being obscured.
Instead of simply pixelating “offensive” images, the processing technology would apply filters to make the image less shocking. The use case explained here is one where a surgical image could be made less repulsive, while still retaining enough detail to understand what was happening.
The core of this particular technology is more about reducing the shocking nature of specific images or footage, rather than making decisions about what is considered offensive or shocking.
That said, the Besançon team built a prototype of a Chrome extension that can automatically identify violence, nudity or medical images and apply visual filters.
Automatic visual censorship: when it's right and when not.
There can certainly be legitimate uses for this AI-based censorship. Protection for kids, social media moderators or police investigators from having to view disturbing images.
Certainly, however, it could also be used to impose unwanted censorship if used improperly or forced on the masses.