🧩 Reporting Harm in AI Companion Chatbots

🤔 What happens when an AI companion crosses a line?

AI companion chatbots increasingly support social, romantic, and intimate interactions. But these interactions can also involve harassment, coercion, manipulation, or other forms of harm.

🔍 What are we studying?

We investigate how companion chatbot platforms support users when something goes wrong.

The project looks at reporting from three perspectives:

1. 📱 Platforms — How are reporting tools designed?

2. 💬 Public discussions — What do users say about harmful experiences and reporting?

3. 🧑‍🤝‍🧑 Users — What makes someone decide to report, or not report, harm?

We also examine what information platforms request and what happens to report data.

💡 So what?

A reporting mechanism is only useful if users understand it, trust it, and feel safe using it.

✨ My interest: identifying barriers to reporting and designing safer, clearer, and more transparent mechanisms for users experiencing harm in AI-mediated relationships.