The artificial-intelligence boom has been difficult for educators. Mainstays of assessment and instruction, such as graded essays and lecture notes, have been hollowed out by chatbots and speech-to-text. Even for students who don’t want to cheat or cut corners, uncomfortable choices abound: do you learn how to use modern tools to get the job done, even if they stop you grasping the underlying principles? Or do you focus on understanding the basics, but fall short on putting theory into practice?

Andrew Ng knows the pain points. As a Stanford University professor, he’s seen first hand the impact of AI on a cohort of undergraduates—and as one of the pioneers of deep learning, perhaps holds some of the responsibility, too. In 2011, he co-founded Google Brain, the search giant’s out-there proposal to use lessons from neuroscience to explore the frontiers of computing. That lab’s early experiments with neural networks eventually led to the transformer, six years later, which unlocked the potential of large language models.

In the years since he also worked at Baidu, a Chinese tech giant, where his group was one of the first to uncover, and name, the “scaling laws”: the discovery that the size of a neural network, the data fed into it and the compute used to train it, have an essentially predictable relationship to the quality of the output. That discovery kicked off the race to raise ever more astronomical funds, which turned AI from an academic curiosity into a big business.

At the same time he was pushing the boundaries of what information technology could do for education. In 2012 he co-founded Coursera, one of the most popular online education platforms, after reasoning that delivering the same lecture in the same hall to a different group of students every year might not be the most efficient way to teach the world about machine learning.

In recent years Mr Ng’s interests in education and AI have begun to coalesce. His latest startup, LearnVector, is only recently operating publicly, but the goal is to use AI’s potential as a tutor to bring the benefits of one-to-one education to more than just the rich. Ideally, it will do so without bringing the disadvantages of AI along for the ride.

I invited Mr Ng on Inside Tech because I wanted to quiz him on how he can remain so optimistic about the impact of AI on education, labour and skills when so many others in his field are uttering warnings about rising unemployment and the inevitable arrival of artificial general intelligence. I also wanted to know, from someone who’s worked in tech giants on both sides of the Pacific, whether he shares the concern of the big AI labs about the increasing capability of Chinese open-source models.

If you’ve found yourself increasingly alarmed by some of the conversations about the technology, then Mr Ng should be a breath of fresh air. Join me on Tuesday at 6pm to find out more. Once you’ve watched the episode, tell us what you think at insider@economist.com.