WSJ Opinion: Free Expression Thinking With Machines
Dec 4, 2025
Vasant Dhar, a data science professor at NYU and author of 'Thinking With Machines,' explores the rapid evolution of AI and its implications. He discusses the balance between leveraging AI as a tool and the risks of overreliance, which could deepen societal divisions. Dhar also highlights the transformative potential of AI in fields like finance and medicine, cautioning about the accuracy of predictions based on domain knowledge. He emphasizes the need for robust governance to ensure accountability while maximizing innovation.
AI Snips
Chapters
Books
Transcript
Episode notes
Common Sense Became Data-Driven
- Modern AI collapsed the boundary between expertise and common sense by learning from vast language data.
- That serendipitous shift made powerful, general-purpose models widely accessible.
From Expert Systems To Deep Learning
- Vasant Dhar traced his AI journey from 1979 expert systems to machine learning on Wall Street and then deep learning.
- He used finance work to illustrate how shifting paradigms resolved past bottlenecks.
Accuracy Depends On Domain And Cost
- Predictive accuracy varies greatly by domain and signal quality; finance is noisy while driverless cars demand near-perfect accuracy.
- The cost of error determines whether we trust AI in a field.
