Understanding the Complexities of Responsibly Sharing NSFW Content Online

2025-11-16

AI Summary

This study addresses the risk of non-consensual sharing of NSFW content on social platforms, revealing how users on sites like Reddit employ implicit linguistic cues—such as euphemisms and contextual hints—to redirect traffic to adult-content platforms (e.g., OnlyFans), thereby facilitating unauthorized dissemination. To overcome the limitations of conventional content moderation systems that rely heavily on explicit keyword filtering, we propose the first systematic taxonomy of non-consensus sharing language patterns in NSFW communities. Building upon RoBERTa, we develop a fine-grained behavioral intent recognition model specifically designed to detect non-consensual sharing intentions. Experimental results demonstrate that our model significantly outperforms baseline methods—including GPT-4, logistic regression, and random forest—in accuracy and robustness. The source code and pre-trained model weights are publicly released, enabling practical deployment of behavior-aware, deep semantic content safety solutions for platform operators.

Technology Category

• Natural Language Processing: Safety and Robustness
• Application Domains: Humanities & Computational Social Science
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Responsible Web: Consent frameworks and practices on the web
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

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, demonstrating superior performance in this specific task. The source code and trained model weights are publicly available at https://github.com/socsys/15NSFW Subreddits.

Problem

• Examining complexities of responsibly sharing NSFW content online
• Identifying linguistic cues linked to non-consensual content sharing behavior
• Developing AI model to detect non-consensual NSFW content sharing

Innovation

• Used RoBERTa-based model for content classification
• Outperformed GPT-4 in identifying non-consensual sharing
• Publicly released source code and model weights