For more than a decade Ng has done both jobs at once: teaching machine learning at scale, first through Coursera and later DeepLearning.AI, while arguing with the industry he helped build. His electricity analogy made his name. Less familiar is the dissent, on risk, regulation and jobs.
1. Why Does Andrew Ng Call AI the New Electricity?
"Just as electricity transformed almost everything 100 years ago, today I actually have a hard time thinking of an industry that I don’t think AI will transform in the next several years."Andrew Ng, Stanford Graduate School of Business, March 2017
The analogy works because electricity was infrastructure. Every industry that adopted it reorganised around it, and Ng expects AI to arrive the same way. Although aggregators reproduce the compressed slogan with no sourced utterance attached, Stanford published the full sentence from his MSx Future Forum talk, delivered while he was chief scientist at Baidu Research. He now runs AI Fund and chairs Landing AI.
2. What Does Andrew Ng's Mars Comparison Say About AI Risk?
"I don’t work on not turning AI evil today for the same reason I don't worry about the problem of overpopulation on the planet Mars."Andrew Ng, The Register, March 2015
Timing is the argument, not denial. Ng separates intelligence from sentience and allows that killer robots could exist someday, while refusing to treat them as today's work. Onstage at Nvidia's GPU Technology Conference, he was answering superintelligence warnings from Stephen Hawking and Elon Musk. Geoffrey Hinton's warnings about AI risk put the same scenario at 10 to 20 percent. A decade on, Ng has not softened it.
3. How Would Andrew Ng Regulate AI Without Harming Open Source?
"It makes more sense to regulate a blender, rather than its motor."Andrew Ng, TIME, August 2024
Regulate the application, not the technology underneath. A motor maker cannot guarantee nobody builds a bomb with it, Ng wrote, so liability should not sit with model developers. His stated worry is the chilling effect on open-weight releases, not safety work itself. SB 1047 cleared the California assembly the day before the op-ed ran, and Newsom vetoed it a month later. Yann LeCun's case for open source AI rests on the same premise.
4. Does Andrew Ng Think Non-Engineers Should Learn to Code?
"I think actually it's time for everyone of every job role to learn to code."Andrew Ng, Y Combinator AI Startup School, June 2025
Yes, and Ng extends it past engineering entirely. His CFO, head of talent, recruiters and front-desk person all know how to code, each performing better for it. Worth naming the interest: he co-founded Coursera and founded DeepLearning.AI, so more technical students is his business. His own hiring is the testable part. Andrej Karpathy on AI-assisted coding argues plain English is now the interface to code, which lowers that bar.
5. Andrew Ng Puts Good Data Ahead of Big Data
"In many industries where giant data sets simply don’t exist, I think the focus has to shift from big data to good data."Andrew Ng, IEEE Spectrum, February 2022
Fifty thoughtfully engineered examples can be enough to teach a neural network what you want it to learn, which is why Ng treats data rather than architecture as the thing worth engineering. Eliza Strickland drew the line out of him for IEEE Spectrum in February 2022, as he named the data-centric AI movement. Landing AI, which he founded, was then selling small-dataset computer vision to manufacturers. It has since repositioned around document processing.
6. Why Does Andrew Ng Say AI Is Both Amazing and Limited?
"The tricky thing about AI is that it is amazing and it is also highly limited... And understanding that balance of how amazing and how limited it is, that’s difficult."Andrew Ng, NBC News, December 2025. Excerpt.
Capability and limitation are not opposites, which is why Ng holds both at once: the systems impress and still fall short of replacing people. NBC ran the interview under exactly that claim. Jared Perlo caught him on the sidelines of his own AI Developers Conference in New York, and the piece landed in December 2025 as the AI-bubble argument peaked. The limits case comes from the founding lead of Google Brain.
7. How Long Does Andrew Ng Expect the Agentic Era to Last?
"I think 10 years from now, we'll still be figuring out new agentic workflows to build."Andrew Ng, Bain & Company, April 2026
A decade at least, on his own estimate. Ng says in the same answer that he wishes it could be done in one or two years, and the consolation he offers is not the timeline but the trajectory: models keep improving, which makes agentic workflows easier and cheaper to build. He spoke to Chuck Whitten, who leads Bain's global Digital practice, on a page where Bain introduces him as "our strategic partner".
8. Andrew Ng Sees an AI Performance Gap Already Open at Work
"I think at this moment in time, we are already seeing a very clear performance gap in many jobs."Andrew Ng, Forbes, May 2025
The gap is present tense. It sits inside jobs people already hold, rather than in a forecast about headcount, and Ng expects it to widen as the models improve. Interviewed at the Imagination in Action event in April 2025, he named coding, decisive in software engineering and now needed outside it, alongside knowing how to prompt large language models. Ng teaches the second himself: our review of Generative AI for Everyone covers the six-hour course.
9. Does Andrew Ng Think AI Will Cause Mass Unemployment?
"There will be no AI jobpocalypse."Andrew Ng, The Batch, DeepLearning.AI, May 2026
No. Ng calls overblown predictions of large-scale unemployment irresponsible and damaging, and his letter cites US unemployment at 4.3 percent alongside software-engineering hiring that held up as coding agents advanced. He also names the incentive: AI companies now price against salaries rather than software seats, which rewards telling employers that headcount is replaceable. His counter-forecast is an "AI jobapalooza". For the other side, see which jobs AI is most likely to reach first.
What to Read Next
Ng is best read against the people he argues with, and alongside the one who largely agrees. Hinton makes the case for the risk Ng sets aside, LeCun is as unconvinced by the extinction case as Ng is, and Kai-Fu Lee treats job loss as the social problem Ng says is overstated.
- Geoffrey Hinton quotes on AI risk: the Nobel laureate who puts an AI takeover at 10 to 20 percent, against Ng's Mars comparison
- Yann LeCun quotes on open source AI: the closest thing Ng has to an ally on open models, and on how far off human-level AI still is
- Kai-Fu Lee quotes on AI and jobs: the displacement case Ng argues against, from the 01.AI chief executive who wrote AI Superpowers
- Will AI take your job? 24 expert quotes: 24 quotes on work and automation, Ng among them, so you can see where his position sits
Cite this page
Aiifi Staff. "9 Best Andrew Ng Quotes on AI." Aiifi, 2026-08-04. https://aiifi.ai/post/andrew-ng-quotes
