Show notes
S02E11 (#321). How do we know when to trust a system? Carol Smith leads the Trust Lab team at Carnagie Mellon Universty, where they conduct research into making trustworthy, human centered, and responsible AI systems. Our conversation highlights the importance of guardrails and ethical considerations in AI development, as well as to ask the right questions and to be critical of the work we are doing – in order to make the best systems we can for the people who are using them or who will be affected by them.
“If the system is providing the right kind of evidence of how it’s making decisions, how it’s making recommendations, if it is a situation where the people understand the capabilities of that system in that particular context, and also know what the edges are – it can’t handle this type of situation, or it will perform poorly in this type of situation – then they can begin to build what is called calibrated trust. “
– Carol Smith
(Listening time: 35 minutes, transcript)
References:
- Full transcript for this episode
- Carol Smith
- LinkedIn: https://www.linkedin.com/in/caroljsmith
- Website: https://www.carologic.com/
- [08:50] How do I help users build and calibrate trust in my product?
- [11:00] Awful AI – a curate list of scary uses of AI
- [13:00] Keeping humans in the loop
- [15:00] Design Risks: How to Assess, Mitigate, and Manage Them
- [15:40] Understanding Data Drift in Machine Learning
- [19:30] Humble Experience Design
- [21:30] Explainability and Trust
- [28:15] Arc Search’s AI responses launched as an unfettered experience with no guardrails
- Recommended listening:
- Episode 270: Design for safety with Eva Penzeymoog
- Sign up to our newsletter – Show notes to your inbox, plus extra stories, news and offers
- Support UX Podcast – Contribute with your time as a volunteer, or make a small financial contribution to help keep the podcast running
This conversation was recorded at UXLx 2023.
