EP15: A Deep Dive into AI Safety - a discussion with Dr. Dvijotham Krishnamurthy on Enhancing Reliability in Language Models
Description
Join us in this enlightening episode as AI researcher Dr. Dvijotham "DJ" Krishnamurthy (Google Deepmind) elucidates methods for increasing the safety and dependability of AI models like GPT.
The conversation probes the concept of instructing these models to sequentially disclose their decision-making processes, with the aim to augment their problem-solving efficiency and promote enhanced collaborations between humans and AI.
The discussion also spans learning methodologies for AI, the enduring importance of human expertise, and the future of AI value alignment through specifications.
Guest:
Dr. Dvijotham Krishnamurthy (https://dj-research.netlify.app/)
Hosts:
Chaithanya (Chait) Bandi - https://www.linkedin.com/in/chaithanya-bandi/
Siddharth (Sid) Shankar - https://www.linkedin.com/in/siddshankar/
For more episodes, subscribe to our podcast at Spotify and Apple:
Spotify: https://tinyurl.com/air-podcast-spotify
Apple: https://tinyurl.com/air-podcast-apple
United States
