Equity and AI in Global Health: Exploring Large Language Models, Building Chatbots and Embracing Discomfort

Episode 37  |  35 minutes

Jonathan Jackson sits down with Brian DeRenzi, Dimagi’s Research and Data team lead, to discuss Dimagi’s work exploring large language models to create chatbots for global health and development use cases. They discuss how we might leverage AI to advance equity despite the reality that it can also decrease equity,  while recognizing the irreplaceable value of human-to-human connection in healthcare. We also discuss the potential of ChatGPT to support more accessible SMS workflows, how voice to text can support non-literate populations, and how to embrace the discomfort of this moment in a way that propels us towards creating positive impact.

Topics include:

• Exploring large language models and GPT-4 for Chatbots in Global Health and Development
• Building tools to support and elevate equity in AI
• AI's Impact on Productivity
• Potential Hype and Pitfalls of AI
• Measuring utility, accuracy, safety, and purpose adherence in chatbots 
• The importance and opportunity in using LLMs with SMS and Voice-to-Text for non-literate populations
• Leveraging AI in a way that elevates and supports human effort instead of replacing it

Related resources:

• GatesNotes, The Age of AI has begun: https://www.gatesnotes.com/The-Age-of-AI-Has-Begun
• Palantir, Bending Artificial Intelligence to our collective will: https://www.palantir.com/newsroom/letters/our-new-platform/
• FarmerChat: https://farmer.chat/
• Comparing Physician and Artificial Intelligence Chatbot Responses to Patient Questions Posted to a Public Social Media Forum:
  https://jamanetwork.com/journals/jamainternalmedicine/article-abstract/2804309