Guest Appearance
Anastassia was invited by Peter M Bach to discuss AI Literacy.
Episode Description
Summary
Most of us now use AI every day. Very few of us could explain what it is actually doing.
Anastassia Lauterbach thinks that gap is the most dangerous thing about this technology, and she has built a company to close it. She is a linguist, a trained clinical psychologist, a former senior executive at Qualcomm and Deutsche Telekom, a former non-executive director at easyJet, Dun & Bradstreet and Wirecard, and now founder and CEO of AI Edutainment in Basel.
In this episode, she draws a line most companies and governments have quietly erased. AI literacy is not AI fluency. Fluency is knowing how to prompt. Literacy is understanding why every large language model hallucinates, where your data goes when you upload it, and when the honest answer is to not use AI at all. Invest only in fluency, she argues, and what you scale is mediocrity.
We work through the four principles of her Responsible AI white paper: changing the architecture itself, radical transparency about failure, literacy as understanding rather than tool training, and the fusion of STEM with the humanities. Along the way, we discuss the elevation of AI into a religion complete with its own popes and heretics, why cognitive capital deserves a place next to human capital on the balance sheet, and what an “Ozempic for the brain” would even look like.
There is also a grand piano, absolute pitch, a childhood in Moscow spent hanging oranges on a Christmas tree, the collapse of Wirecard, and a ginger cat who serves as Chief Marketing Officer.
Chapters
- 0:00 Intro
- 1:59 Guest Intro, Basel and Piano
- 15:15 Anastassia’s early journey
- 21:43 The Brain and critique on today’s AI and society
- 35:49 Anastassia’s Mission in AI Literacy
- 42:34 AI Literacy vs. AI Fluency
- 48:08 The cognitive challenge and appetite for understanding
- 51:52 Mediocrity and the need to celebrate human effort
- 55:18 Education reforms and re-exercising our brain
- 1:09:37 Reaching the four pillars of Responsible AI
- 1:19:17 Anastassia’s Vision
- 1:24:05 Q&A Start
- 1:24:32 What does innovation mean to you?
- 1:27:14 Key moment, event, person
- 1:28:22 Time Management
- 1:32:49 Favourite childhood memory
- 1:35:16 Biggest challenge to date
- 1:41:34 Advice for young professionals
- 1:43:38 What would you most like to be remembered for?
- 1:43:52 Where can people find you?
- 1:46:39 Final Message
- 1:46:58 Outro
Key Takeaways
- Literacy is not fluency. Fluency is prompting. Literacy is understanding what a model can and cannot do. Almost every corporate and government AI programme teaches the second and skips the first.
- Hallucinations are structural, not a bug. They cannot be engineered out of transformers. What can be built is better governance, better architecture and better human judgement around them.
- Four principles of responsible AI: change the architecture (causality, world models, theory of mind), report failures as radically as successes, teach literacy rather than tools, and fuse STEM with the humanities.
- Cognitive capital. If human capital sits on the balance sheet, why not the cognitive capacity we are quietly offloading to machines?
- A life in structures. From absolute pitch and a Moscow music school to Qualcomm, the Wirecard collapse, and rebuilding from nothing with a cat as Chief Marketing Officer.