I believe their point is that surveillance systems that are used to identify people are being trained on social media images and they don’t want those systems to be better. The point is not personal risk, but systemic/societal risk