Too many digital mental health startups are taking an ML/AI-first approach to product development. Before that, we need a better foundation of non-ML/AI data science in the digital mental health space.
Wysa's anonymity guarantee and focus on developing a system that actually has emotional intelligence make me excited for what they'll accomplish. Their design choices make me wonder if a more Rogerian than CBT approach might be the secret to emotional intelligence in a therapybot.
Woebot offers useful CBT coaching. But its chatbot interface is more distracting than helpful at this point, reflecting AI's current inability to understand natural language.
The designers of this app have made some great choices around mood tracking and introducing core CBT skills. But they've also made some questionable choices around data capture and analytics.
Part 2 of an exploratory series on psychotherapeutic chatbots that include AI capabilities. How effective is self-guided Internet-based psychotherapy?
Part 1 of a series on psychotherapeutic chatbots that include AI capabilities. How effective is one-on-one psychotherapy?
Will AI-based chatbots do any better than earlier incarnations of electronic psychotherapy?