The point was to find out if malicious actors could use such tools for convincing interactions, and although I do disagree with their approach, I think that it

The point was to find out if malicious actors could use such tools for convincing interactions, and although I do disagree with their approach, I think that it would be impossible to conduct such a experiment with user consent on a subreddit scale.