Last Week in AI Therapy Chatbots, AI Ethics at Google, and Higher-Res Climate Data
Jul 19, 2020
Delve into the fascinating world of therapy chatbots like Wobot, designed to make mental health support more accessible during the pandemic. Explore their potential and limitations from a clinical perspective, emphasizing the need for human touch in therapy. Learn about Google's efforts to ensure ethical AI practices amidst innovation. Finally, uncover how generative adversarial networks are enhancing climate data, while maintaining a critical view on their reliability, and the ongoing debates over facial recognition technology and its societal impacts.
AI Snips
Chapters
Transcript
Episode notes
Woebot Experience
- Andrey tried Woebot, an AI chatbot for mental health.
- He found the experience bizarre and wrote an article about it.
Woebot's Approach
- Woebot uses established therapeutic techniques like CBT.
- This approach makes it more effective than open-ended AI chatbots.
Chatbot Imperfections
- Therapy chatbots aren't perfect and can make inappropriate responses.
- Sharon's experience with Replica illustrates this, as she stopped using it after a day.