Start here
The free tool below runs entirely in this browser — no account, nothing stored or transmitted. Redact anything you would not want kept before you type it.
Open the companion tool →How language models simulate emotional concern
Large language models predict empathetic words based on patterns in human literature and conversation datasets.
Recognizing that an AI has no feelings, consciousness, or lived experience helps maintain healthy expectations.
The protocol
- Acknowledge that feeling comforted by well-written text is a natural human reaction.
- Remind yourself that the chatbot's words are mathematical predictions, not personal feelings.
- Identify 2-3 real people in your life or community with whom you can share mutual experiences.
- Use chatbots for writing or organization while keeping emotional vulnerability grounded in human relationships.
What the tool does
Explore reflection prompts on emotional processing and human relationships. Runs in browser.
Limit first: This is a self-reflection tool. It is not mental health treatment or professional therapy.
Open the companion tool →One human next step
If you are experiencing grief, sadness, or isolation, reach out to a trusted professional or call/text 988 (US) for free, confidential support.
Research log and safety checks
AI-assist path: reflection builder. §6.10: not a frontier-pace topic.
§0.5 protection result: category is chatbot-dependency-check; crisis-adjacent: no; fear/urgency/scarcity toolkit used: no. The protective function is free and the paid feature is convenience only.