This is a bit within the realm of my research, so I feel confident about commenting: yes, it's possible. There might be a catch-22 in that the people who would

This is a bit within the realm of my research, so I feel confident about commenting: yes, it's possible. There might be a catch-22 in that the people who would be most resilient are probably those who would be least concerned about the welfare of victims. However, I don't know that's really perfectly accurate: I think you could find some medium, or identify those who empathize but tend not to become easily upset. Empathy and emotional response are kind of dissociable. Interestingly, the rates of stress disorder in those populations that they cite in the article are kind of in the ballpark of what you see in epidemiological studies of trauma-exposed populations, like combat exposed populations, areas subject to frequent terrorism, natural disaster survivors and so forth. My impression based on what they're seeing is that it's comparable to a job with high trauma exposure rates. It might be a prime opportunity to study trauma response and treatment, actually, in that often these things have an element of unpredictability to them. If content moderation is something that's reliably characterized by traumatic exposure, you could maybe try to study content moderators to identify predictors of and preventions for stress response.