Driverless Cars Are Biased Against Dark-Skinned People and Kids, Scientists Say
futurism.com
futurism.com
I have worked in the video analytics field and in particular traffic control and pedestrian detection for several years and although it requires a greater amount of false positives to cover the detection of children the standard is to error on the side of caution.
As far as the pigment of peoples skin, this is hogwash as the amount of skin exposed compared to clothing is minimal. Was this study conducted using nude pedestrians? I can assure everyone that the detection of a person wearing all black and a hoodie(Aka Antifa uniform) is the minimum requirement.
Many detection systems use dual cameras and infrared and thermal imaging, this seems to be either propaganda or "Scientists" finding data to support their own bias.
Not only, the scientists had no access to the actual systems (and sensors) used in the actual driverless cars, and assumed (declaring to be certain about that) that they will have the same issues as the systems they tested.
Of course the engineers and programmers working at these driverless cars systems may be a bunch of overpaid good-for-nothing people that only re-use open source software, but it is at least possible that they are instead good at what they do and that these systems are way ahead of the programs and hardware that were tested by the scientists.