A Beginner's Guide to AI

chevron_right

Can You Trust Your AI? Vasant Dhar on Robot Taxis vs. Robot Doctors

Feb 2, 2026

Guest

Vasant Dhar

Vasant Dhar, NYU Stern professor and author focused on decision intelligence and AI governance, outlines when machines should assist and when they must be constrained. He compares low-risk automation like finance to high-stakes areas such as healthcare and mental health. He explains trust as error rates versus consequences, explores edge cases for self-driving cars, and warns about agentic AI and governance challenges.

54:05

forum

Ask episode

web_stories

AI Snips

view_agenda

Chapters

menu_book

Books

auto_awesome

Transcript

info_circle

Episode notes

question_answer

ANECDOTE

Early Medical AI Sparked A Career

Vasant Dhar describes seeing an early medical diagnostic system at Stanford in 1979 that asked diagnostic questions.
That experience sparked his lifelong fascination with machines that can reason about problems.

volunteer_activism

ADVICE

Structure Medical Data And Keep Humans Involved

Use LLMs to structure existing medical data into mineable formats for better decisions.
Keep humans in the loop for high-empathy domains like mental health to avoid harm.

insights

INSIGHT

Trust Depends On Errors And Consequences

Trust in AI depends on mistake frequency and the consequences of those mistakes.
Vasant Dhar maps trust using a heat map balancing error rates and error costs.

Get the Snipd Podcast app to discover more snips from this episode

Vasant's AI origin story and mentors

02:43 • 4min

chevron_right

Why 'Brave New World' as a warning

07:11 • 2min

chevron_right

AI optimism in healthcare versus risks

09:30 • 3min

chevron_right

Data exists but isn't usable in medicine

12:31 • 2min

chevron_right

The trust heat map: errors and consequences

14:31 • 5min

chevron_right

Edge cases and how vehicles handle them

20:00 • 4min

chevron_right

Bounded rationality and machine attention

24:00 • 2min

chevron_right

Where machines outperform humans

25:37 • 3min

chevron_right

Future of work: AI as amplifier

28:29 • 6min

chevron_right

Thinking with machines, not just about them

34:26 • 4min

chevron_right

Governance, obligations and agentic AI

Risks of uncontrolled agents and ethics

43:15 • 4min

chevron_right

AI wow moments and integrated senses

47:07 • 3min

chevron_right

Where to find the book and newsletter

• Mentioned in 28 episodes

Death of a Salesman

Arthur Miller

Written by Arthur Miller, 'Death of a Salesman' is a two-act tragedy set in late 1940s Brooklyn. The play revolves around Willy Loman, an aging salesman who grapples with the reality that his American Dream of success and wealth is unattainable. Through a montage of memories, dreams, and arguments, the play explores Willy's consciousness and his inner conflict. Willy's life is marked by poor choices, delusions, and a desperate attempt to secure a better future for his sons through his life insurance policy. The play critiques the American Dream and the superficial values that Willy has adopted, leading to his tragic downfall. The final scene depicts Willy's funeral, attended by only a few people, highlighting the contrast between Willy's illusions and the harsh reality of his life.

#4808

• Mentioned in 10 episodes

Thinking with Machines

Vasantasant Dar

string

#41

• Mentioned in 271 episodes

Brave New World

Aldous Huxley

Written in 1931 and published in 1932, 'Brave New World' is a dystopian novel set in a futuristic World State where humans are bred in vitro and predestined into a rigid social hierarchy. The novel explores themes of technological advancement, social conditioning, and the loss of individuality and freedom. The story revolves around characters such as Bernard Marx, an Alpha who questions the society, Mustapha Mond, the World Controller, and John, a 'Savage' from a reservation who introduces elements of emotion and individuality into the highly controlled World State. The novel critiques the optimism of technological progress and its potential to create a society that values stability and happiness over humanity and freedom.

🤖🧠 Thinking with Machines with Vasant Dhar

What happens when AI stops being a tool and starts becoming a collaborator and an agent? In this episode, NYU Stern professor and AI pioneer Vasant Dhar takes us through the real story behind modern AI, and the practical frameworks we need for AI trust, AI governance, and the coming era of agentic AI.

🚀 What you will learn

- Why “thinking with machines” is a bigger idea than “thinking machines”

- How the automation frontier separates low-risk automation from high-stakes human control

- Why healthcare has lots of data but still struggles to make good decisions

- Why mental health is a dangerous place to outsource empathy to machines

- What edge cases in AI mean and why they matter for self-driving cars

- How AI agents change the governance conversation, from obligations to restrictions to rights

📌 Key highlights

- A practical definition of trust in AI based on error rates and consequences

- AI in healthcare data: turning medical trails into usable decision intelligence

- The future of work: AI as an amplifier, not a substitute, unless you let it become a crutch

- Governance questions that no one gets to avoid once agents can act in the world

📧💌📧

Tune in to get my thoughts and all episodes, don't forget to ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠subscribe to our Newsletter⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠: ⁠⁠⁠⁠beginnersguide.nl⁠⁠⁠⁠

📧💌📧

About Dietmar Fischer:

Dietmar is a podcaster and AI marketer from Berlin. If you want to know how to get your AI or your digital marketing going, just contact him at argoberlin.com

Quotes from the Episode 💬

“Trust depends on how often a machine makes mistakes and the consequences of those mistakes.”

“In physical health, I’m very optimistic. In mental health, not so.”

“It’ll likely lead to a bifurcation of humanity… skills get amplified… or people rely on the machine as a crutch.”

Chapters ⏱️

00:00 Vasant Dhar’s origin story in AI and early expert systems

05:08 A Brave New World warning and why optimism still needs guardrails

07:26 AI in healthcare vs mental health and why feelings change the rules

12:37 The trust heat map and the automation frontier in real life

18:21 Edge cases, bounded rationality, and what machines pay attention to

26:03 The future of work and why AI amplifies both skill and decline

36:23 Governance, AI agents, and how much agency we should allow

44:05 AI wow moments and the next frontier: integrated machine senses

47:15 Where to find the book, podcast, and newsletter

Where to find Vasant Dhar 🔎

- Visit Vasant's Website, also to find all the links to shops with "Thinking with Machines", his book: vasantdhar.com

- Listen to his Podcast: bravenewpodcast.com

- and get his Newsletter: vasantdhar.substack.com

Music credit: "Modern Situations" by Unicorn Heads`

Hosted on Acast. See acast.com/privacy for more information.

Home Top podcasts Popular guests Top books