Scrutability Limits and Misapplied 'Open Source' Rhetoric

It’s our second week of playing AI lingo bingo. The summit in India is underway and the air is thick with vague terms that fail to describe the big problems.

More like this: Lingo Bingo at the India AI Summit w/ Karen Hao, Joan Kinyua, Chenai Chair, and Rafael Grohmann

With us this week to discuss co-opted terms is Meredith Whittaker on how ‘open source’ cannot meaninfully be applied to AI systems; Audrey Tang on ‘democratisation’, something which is both helped and harmed by AI; Abeba Birhane on everyone’s favourite slogan ‘AI for Good’; and Usha Ramanathan to discuss ‘AI and development’ in the context of the Aadhaar project in India.

Further reading & resources:

• More on Usha Ramanathan — legal researcher and human rights activist
• More on Abeba Birhane — principle investigator at the AI Accountability Lab at Trinity College Dublin
• More on Meredith Whittaker — President of Signal
• More on Audrey Tang — Taiwan’s first Digital Minister
• Distributional AGI Safety — by Nenad Tomašev et al
• Watch this week’s interviews in full on Youtube

**Subscribe to our newsletter to get more stuff than just a podcast — we run events and do other work that you will definitely be interested in!**

Pre Production by Georgia Iacovou | Post Production by Sarah Myles