Exciting news!!! I figured out how to integrate the cognitive restructuring exercise into my chatbot! I’ve also adjusted my bot to address the issues raised in previous testings.
Preliminary user feedback indicated a significant improvement to user’s experience with the bot. In addition to being empathetic to the user, the bot is now giving concrete insights and feedback to the user’s issues. Most users also rated the conversation’s “robotic-ness” between 2-4 (1 = indistinguishable from a human conversation, 10 = very robotic conversation).
There are some issues that needs to be adjusted. First, there is not a lot of flexibility to the bot’s responses. For example, a user cannot ask the bot for clarification on what it said. Second, the there is an approximately 30% chance the bot may give an undesirable output which compromises the flow of the conversation. I aim to address these issues by working with Rejoan Ahmed to create a chat engine that offers more flexibility.
The bot is currently deployed online. If you have experience with LLMs or cognitive behavioral, or if you’re just interested in trying it out, don’t hesitate to send me a message 🙂
(Below are excerpts from a conversation with the bot)
[easy-image-collage id=2017]