Understanding the Complexities of Responsibly Sharing NSFW Content Online

TL;DR

The study analyzes 15 large restricted NSFW subreddits over nine years to understand how NSFW content circulates, monetizes, and potentially involves non-consensual sharing. It reveals that Reddit NSFW spaces often act as gateways to other platforms (e.g., OnlyFans, Kik) and to off-platform trading, with substantial cross-domain link sharing and coordinated promotion. A RoBERTa-based classifier is developed to detect non-consensual sharing, achieving high recall and aiding moderators, while ethics and reporting protocols are documented. The findings have implications for platform policy, moderation strategies, and privacy considerations as NSFW content becomes more mainstream. The work also provides publicly available code and data-oriented tooling to support ongoing moderation efforts.

Abstract

Reddit is in the minority of mainstream social platforms that permit posting content that may be considered to be at the edge of what is permissible, including so-called Not Safe For Work (NSFW) content. However, NSFW is becoming more common on mainstream platforms, with X now allowing such material. We examine the top 15 NSFW-restricted subreddits by size to explore the complexities of responsibly sharing adult content, aiming to balance ethical and legal considerations with monetization opportunities. We find that users often use NSFW subreddits as a social springboard, redirecting readers to private or specialized adult social platforms such as Telegram, Kik or OnlyFans for further interactions. They also directly negotiate image "trades" through credit cards or payment platforms such as PayPal, Bitcoin or Venmo. Disturbingly, we also find linguistic cues linked to non-consensual content sharing. To help platforms moderate such behavior, we trained a RoBERTa-based classification model, which outperforms GPT-4 and traditional classifiers such as logistic regression and random forest in identifying non-consensual content sharing, showing better performance in this specific task. The source code and model weights are publicly available at https://github.com/socsys/15NSFWsubreddits.

Figures (7)

• Figure 1: Locations of posters to Top 15 restricted NSFW subreddits studied in this paper. ($\approx$ 155 countries included)
• Figure 2: Year-wise distribution of number of posts in subreddits
• Figure 3: Social media engagement trends in NSFW subreddits from 2016--24. (a) compares the overall percentage of posts mentioning user handles on social media. (b) reveals the top 15 platforms mentioned. An asterisk indicates public platforms like Instagram and Twitter, often used for public feeds, in contrast to private platforms like WhatsApp or Telegram.
• Figure 4: Illustrate the distribution, authorship, and temporal dynamics of social media username mentions in Reddit posts. (a) shows that while most usernames are rarely repeated, a small fraction dominates the space. (b) reveals a similar inequality in subreddit participation, with a small user base contributing the majority of activity. (c) suggests that high-frequency usernames are often posted by many different users, hinting at viral or coordinated behavior. (d) supports this further by showing compressed posting time windows for these high-frequency usernames—consistent with campaign-driven or automated posting patterns.
• Figure 5: Top 10 Social Media Platforms with Price Mentions (Annotated with Price Range and Inter Quartile Range (IQR))