Photos and videos are harder to spy on.

However, this might be changing too...

https://www.wsj.com/articles/chinas-stopchat-censors-can-now...

Text can by monitored using automated tools, however the same capability is not as well developed for video/photos.

Also, pics/videos have more impact. If I describe the results of a beating it will not have as much of an impact as the photo of the broken bleeding person will have.

So for a government that is trying to control the narrative the ability to control the images being transferred becomes key.