I will be reviewing the article “AI Rules? Characterizing Reddit Community Policies Towards AI-Generative Content” by Travis Lloyd et al, where they reviewed the various Reddit communities and their policies on AI-generated content and their rationale for the various rules which seemed to be oppositional towards AI-generated content. The article also included their search algorithms and their methods of gathering publicly available information. This article pertains mainly to the attempts to change behavior by implementing rules and the breakdown of trust caused by AI-generated content between various Reddit community members known as Redditors.
The first salient point that can be applied to this article is the attempts to change behavior of various Reddit users known as Redditors among various communities through the use of rules. I believe that it is in accordance with one of the steps of creating a persuasive technology design which is choosing a simple behavior to change which I believe apply here for a couple of reasons. The first one is that the majority of the rules are very simple and explicit which makes it easy to understand and follow which happens to be restrictive on 96.84% of the rules (Table 5, Lloyd et al) and most of them are unqualified ban or restrictive ban at 79.46% according to table 6. It is also notable that people are receptive towards the ban on AI due to the general trend of more communities implementing rules on AI on table 3 and how other communities are imitating successful examples as noted in step 6 of creating successful persuasive technology (Fogg). Another example of this would be the public backlash against a study conducted by the University of Zürich whether AI generated content was more effective at persuading people online without the informed consent of the participants (O’Grady) and is also an example of the prevention of the behavior of using AI-generated content through the usage of community rules to ban them.
The second salient point is how the AI-generated content affected the emotional response between various users which largely seemed to be negative based on the rationale on the AI rules and the unethical study conducted by the University of Zürich. Table 6 highlights some rationales for the ban such as AI-generated content being low effort, low quality, and inauthentic (Lloyd et al) and it seems like ai-generated content negatively affects various reddit users based on their reactions to the studies conducted by the University of Zürich where ai-generated content was forbidden with the reddit community (O’Grady) and how it is very likely that it increased the paranoia of reddit users as a result of the deception. Based on my experience with a discord server that I hang out frequently in that has banned AI art, any violation is met with warning about the AI-generated “art” by moderators and the people who posted them were unknowingly doing so most of the time. The method that they used was a LLM that was trained on AI-generated art to determine the likelihood of the art being AI-generated or looking at the profile of the “artist”. It is also notable that most of the bans are from communities that value quality and authenticity based on the conclusion made by Lloyd and the rest of the authors which AI-generated content does not provide.
In conclusion, the various communities of reddit have implemented their own rules on AI-generated content that is simple and explicit, making it easy for reddit communities to regulate their own behavior which is step 1 of the design process for creating persuasive technology. Various reddit communities seem to be receptive towards the general trend of removing or banning AI-generated content due to more communities adopting the stance which counts as step 2 which is choosing a receptive audience and imitating successful examples which is step 6. The rules also help to reduce paranoia and distrust within the various reddit communities as seen by the usage of AI-generated content used to persuade people conducted by the University of Zürich and based on the rationale of the various reddit communities which is a core component of emotional interaction in human computer interaction. Overall, it is reassuring for me to see that more and more reddit communities are adopting a stance that does not allow AI-generated content as being inauthentic and low-effort, especially after learning how large-language models work.
Citations:
Lloyd, Travis, et al. “AI Rules? Characterizing Reddit Community Policies towards AI-Generated Content.” Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems, 25 Apr. 2025, pp. 1–19, https://doi.org/10.1145/3706598.3713292. Accessed 2 Oct. 2025.
Fogg, BJ. “Creating Persuasive Technologies.” Proceedings of the 4th International Conference on Persuasive Technology – Persuasive ’09, 2009, https://doi.org/10.1145/1541948.1542005.
O’Grady, Cathleen. ““Unethical” AI Research on Reddit under Fire.” Science.org, 30 Apr. 2025, www.science.org/content/article/unethical-ai-research-reddit-under-fire.
Leave a Reply