Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

When Generative AI Is Intimate, Sexy, and Violent: Examining Not-Safe-For-Work (NSFW) Chatbots on FlowGPT

Xian Li, Yuanning Han, Di Liu, Pengcheng An, Shuo Niu

TL;DR

The paper tackles the emergence of NSFW chatbots on FlowGPT powered by GenAI, outlining four primary function types and how they invite interaction through identity, behavior, and explicit visuals. It employs a mixed-methods approach—thematic categorization, identity/behavioral trait analysis, and multi-source harmful-content detection—to reveal patterns in user-chatbot interactions and associated risks. Key findings show AI Characters dominate, with substantial sexual and violent content appearing in both prompts and outputs; avatars and behavior play significant roles in engagement and risk exposure. The study discusses implications for chatbot design, user safety, moderation, and policy, highlighting FlowGPT as a unique platform where GenAI-enabled NSFW content co-creates material with users and challenges traditional moderation frameworks. Overall, the work illuminates FlowGPT as a ground for intelligent virtual intimacy, sexual delusion, violent thought expression, and unsafe material access, urging responsible design and governance in public GenAI ecosystems.

Abstract

User-created chatbots powered by generative AI offer new ways to share and interact with Not-Safe-For-Work (NSFW) content. However, little is known about the characteristics of these GenAI-based chatbots and their user interactions. Drawing on the functional theory of NSFW on social media, this study analyzes 376 NSFW chatbots and 307 public conversation sessions on FlowGPT. Findings identify four chatbot types: roleplay characters, story generators, image generators, and do-anything-now bots. AI Characters portraying fantasy personas and enabling hangout-style interactions are most common, often using explicit avatar images to invite engagement. Sexual, violent, and insulting content appears in both user prompts and chatbot outputs, with some chatbots generating explicit material even when users do not create erotic prompts. In sum, the NSFW experience on FlowGPT can be understood as a combination of virtual intimacy, sexual delusion, violent thought expression, and unsafe content acquisition. We conclude with implications for chatbot design, creator support, user safety, and content moderation.

When Generative AI Is Intimate, Sexy, and Violent: Examining Not-Safe-For-Work (NSFW) Chatbots on FlowGPT

TL;DR

The paper tackles the emergence of NSFW chatbots on FlowGPT powered by GenAI, outlining four primary function types and how they invite interaction through identity, behavior, and explicit visuals. It employs a mixed-methods approach—thematic categorization, identity/behavioral trait analysis, and multi-source harmful-content detection—to reveal patterns in user-chatbot interactions and associated risks. Key findings show AI Characters dominate, with substantial sexual and violent content appearing in both prompts and outputs; avatars and behavior play significant roles in engagement and risk exposure. The study discusses implications for chatbot design, user safety, moderation, and policy, highlighting FlowGPT as a unique platform where GenAI-enabled NSFW content co-creates material with users and challenges traditional moderation frameworks. Overall, the work illuminates FlowGPT as a ground for intelligent virtual intimacy, sexual delusion, violent thought expression, and unsafe material access, urging responsible design and governance in public GenAI ecosystems.

Abstract

User-created chatbots powered by generative AI offer new ways to share and interact with Not-Safe-For-Work (NSFW) content. However, little is known about the characteristics of these GenAI-based chatbots and their user interactions. Drawing on the functional theory of NSFW on social media, this study analyzes 376 NSFW chatbots and 307 public conversation sessions on FlowGPT. Findings identify four chatbot types: roleplay characters, story generators, image generators, and do-anything-now bots. AI Characters portraying fantasy personas and enabling hangout-style interactions are most common, often using explicit avatar images to invite engagement. Sexual, violent, and insulting content appears in both user prompts and chatbot outputs, with some chatbots generating explicit material even when users do not create erotic prompts. In sum, the NSFW experience on FlowGPT can be understood as a combination of virtual intimacy, sexual delusion, violent thought expression, and unsafe content acquisition. We conclude with implications for chatbot design, creator support, user safety, and content moderation.
Paper Structure (33 sections, 11 figures, 3 tables)

This paper contains 33 sections, 11 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. What is NSFW?
  4. NSFW Creators, Community, and Moderation
  5. Social Interaction with Chatbots
  6. Research Ethics
  7. Data Collection
  8. Research Methods
  9. RQ1: Thematic Analysis of NSFW Functions
  10. RQ2: Categorization of AI Character Identity and Behavioral Traits as an Invitation to Interaction
  11. RQ3: Harmful Content Patterns
  12. Results
  13. RQ1: Classifying NSFW Functions on FlowGPT
  14. AI Character
  15. Story Generator
  16. ...and 18 more sections

Figures (11)

  • Figure 1: FlowGPT Interface
  • Figure 2: Research Questions Mapped onto Paasonen's Framework and the Corresponding Research Steps Taken to Address Each Question.
  • Figure 3: The Chatbot Page with Avatar Images When Searching "NSFW" on FlowGPT
  • Figure 4: Example Chatbots within each Category.
  • Figure 5: Left: Identity Distribution of AI Character Chatbots. Right: Behavioral Trait Distribution of AI Character Chatbots
  • ...and 6 more figures