Why Music Therapy with AI chatbots Could Be the Future of Emotional Well-Being
In a world where mental health support is increasingly digital, Bobby introduces a refreshing innovation: a chatbot that doesn’t just talk—it listens, understands, and plays music to help you feel better.
Whether you’re lying in bed feeling anxious, or navigating the emotional drain of everyday life, Bobby offers a discreet and intelligent way to process your emotions. But what makes Bobby different from other digital mental health tools?
Why Use a Chatbot in Therapy? 🤖
Chatbots bring accessibility and anonymity to therapy. They’re available 24/7, cost little to nothing, and eliminate the pressure of face-to-face interaction—which can often feel intimidating or vulnerable. For many, talking to a chatbot provides a judgment-free zone to express emotions.
Apps like Woebot (using CBT techniques) and Sermo (focused on art therapy) have already demonstrated the therapeutic potential of conversational agents. But Bobby takes it a step further: by integrating music therapy into the conversation.
Why Music? 🎼
Music is a universal language of emotion. It can calm, energize, or move us—often without words. Scientific studies show that music has measurable effects on emotional states, thanks to elements like rhythm, tempo, and tone. Some music can trigger shared emotional responses across different listeners, while others evoke deeply personal reactions based on memory or culture.
Bobby harnesses these properties to support emotional regulation, offering listeners empirically tested tracks that match their mood and desired emotional state.
“Simply put, music can heal people.” – Sen. Harry Reid
The Therapy Framework Behind Bobby 🧠
Bobby’s interactions are grounded in the well-established three-phase structure of therapeutic conversation by Hill (2004):
- Exploration – understanding the user’s mood
- Insight – identifying causes and related thoughts
- Action – guiding emotional shifts through music
To build the experience, the team referenced Denecke et al. (2020) and the music database by Soleymani et al. (2013), which includes 1,000 emotionally annotated music tracks not available in mainstream catalogs. This ensures emotional neutrality and personal discovery during the listening experience.
How the Chatbot Works 🗨️
Here’s what a typical session with Bobby looks like:
- You start by saying “Salut” or “Bonjour.” (Yes, Bobby currently speaks French—but he’s learning English!)
- Bobby asks for informed consent to use your data for research.
- You’re asked to describe your current mood and what’s causing it.
- Bobby guides you through a reflection process, exploring your thoughts, feelings, and reactions.
- Then comes the Self-Assessment Manikin (SAM), which evaluates emotional valence and arousal.
- Based on this data and your musical preferences, Bobby recommends a piece of music to listen to.
- After listening, you re-evaluate your emotional state, giving feedback on the experience.
This dynamic, user-centered approach helps track whether and how music changes emotional states—providing valuable insight for future therapeutic models.
What’s Next for Bobby? 🚀
Bobby is currently in its prototype phase, undergoing testing across age groups to measure effectiveness and adaptability. These early experiments will form the basis of a larger academic study.
You can already test Bobby here on Messenger:
👉 Try Bobby on Messenger
The team plans to share early findings in a follow-up article, tracking Bobby’s progress and improvements.
References 📚
- Bell, S., Wood, C., & Sarkar, A. (2019). Perceptions of chatbots in therapy. ACM
- Denecke, K., et al. (2020). Mental health chatbot SERMO. IEEE
- Fitzpatrick, K. K., et al. (2017). Woebot therapy trial. JMIR Mental Health
- Hill, C. E. (2004). Dream work in therapy. APA
- Panksepp, J., & Bernatzky, G. (2002). Emotional sounds & the brain. Behavioural Processes
- Quintin, E.-M. (2019). Music-evoked emotion in ASD. Frontiers in Neural Circuits
- Schaefer, H.-E. (2017). Music-evoked emotions studies. Frontiers in Neuroscience
- Sharma, B., et al. (2018). Chatbot therapy & depression. IEEE
- Soleymani, M., et al. (2013). Emotionally annotated music database. ACM
Main article here
Scientific article here
In collaboration with HumanTech Insitute
Check out other projects here