Last Week in AI Bad Uses of AI, Google and Margaret Mitchell, AI for Fairer Healthcare
Jan 28, 2021
The podcast dives into the ethical challenges surrounding AI, spotlighting Google’s internal struggles after sidelining a key researcher. It examines a controversial South Korean chatbot and a faulty app claiming to assess trustworthiness. The conversation then shifts to the biases in pain assessment in healthcare, proposing an AI approach that blends patient reports with objective data for fairer evaluations. Lastly, it explores the potential of AI to revolutionize medical practices while emphasizing the importance of careful implementation.
AI Snips
Chapters
Transcript
Episode notes
Google Locks Out Researcher
- Google locked Margaret Mitchell out of her corporate account, raising concerns.
- Mitchell, an AI ethics researcher, was investigated after allegedly looking for evidence related to Timnit Gebru's firing.
Chatbot Controversy
- ScatterLab's chatbot, Lee Luda, trained on billions of Korean conversations, sparked controversy.
- Luda spewed hate speech, highlighting the risks of open-domain conversational AI.
Trustworthiness App
- DeepScore, a Tokyo-based company, developed an app claiming to detect trustworthiness.
- The app analyzes facial movements and voice, raising ethical concerns about its accuracy and potential biases.