The media have an effect on the direction of politics through the intensity and frequency of their coverage. Likewise omissions, self-censorship, and neglect have effects.
The internet has an image of Jimmy Sham, a civil rights supporter in Hong Kong, following a hammer attack. The image shows a bloodied man on blood splattered concrete. I searched the Washington Post for the man's name. I found 4 articles posted within the past few hours. The articles feature images and video from the Hong Kong legislature but no image of the bloodied man. It seems the 2019 story of the Hong Kong protests is essentially about the direction of governance and violence. On most days violent conflict is the essence of the story.
The relevant question is newsworthiness. It may be sufficient from a newsworthiness perspective to write that the man has bloody head injuries but if that is the standard then it would seem that newspapers do not need to have images at all. The paper can choose to report that some legislators protested the Chief Executive without choosing to show images of those protestors doing so by holding banners and heckling the speaker while standing on their desks.
It seems the salience of both stories, that of the protesting legislators and that of the bloodied civil rights supporter, is reduced if images are omitted. The decision over a bloody image at a particular newspaper on a particular story is not nearly as important as whether there is a systematic problem. Is there an systematic problem of self-censorship of images of violence? In the modern era a newspaper is received on a computer screen and the technology exists to blur the image until the user clicks to consent to see the proper image.
Addendum: I just discovered the SCMP (South China Morning Post) has a pixelated version of the image on their website. This makes me wonder if there is a U.S.-centric problem and whether it need be pixelated at all in any newspaper.