We want accurate AI, right? As long as it’s accurate, we’re all good? My guest, Will Landecker, CEO Accountable Algorithm, explains why accuracy is just one metric among many to aim for. In fact, we have to make tradeoffs across things like accuracy, relevance, and normative (including ethical) considerations in order to get a usable model. We also cover whether explainability is important and whether it’s even on the menu and the risks of multi-agentic AI systems.
Will Landecker is a technical consultant for responsible AI. He has a PhD in machine learning and over 10 years of experience working at tech companies like Twitter, Lyft, Stripe, and Nextdoor.
Website: https://www.accountablealgorithm.com/
Subscribe where you listen:
Spotify | Apple Podcasts | iHeart Radio | PocketCast | YouTube
Reid Blackman is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.
Share this post