From the statement by the researchers

While we did not write any comments ourselves, we manually reviewed each comment posted to ensure they were not harmful

From the post

Some high-level examples of how AI was deployed include:

AI pretending to be a victim of*removed*
AI acting as a trauma counselor specializing in abuse
AI accusing members of a religious group of "caus[ing] the deaths of hundreds of innocent traders and farmers and villagers."
AI posing as a black man opposed to Black Lives Matter
AI posing as a person who received substandard care in a foreign hospital

They also used another LLM to scrape data about the op in each topic they replied to. Sick