Introduction

Questions around the possibilities and potential dangers of Artificial Intelligence cover the headlines these days, but are these actually new questions?

Computer scientist Peter Norvig has been writing about AI and the ethics of data science for years. Before he was a professor at Stanford University’s Human-Centered Artificial Intelligence Institute, he worked for NASA and held a major consulting role at Google. His books, Artificial Intelligence: A Modern Approach (4th Edition) and Data Science in Context: Foundations, Challenges, Opportunities, explore the theory and practice of AI and data science.

Peter and Greg discuss the cyclical nature of new technology mania, the misconceptions of modern AI,  and the different ways companies could monetize these systems in the future. 

*unSILOed Podcast is produced by University FM.*

Episode Quotes:

Open source and AI Systems

27:56: One reason to open source is if you have a vibrant open-source community, it's hard for one individual company to compete against that. One of the places I worked was Sun Microsystems. They had their own version of Unix. But that wasn't sustainable. You know, one company couldn't compete against the entire open-source Linux community. And I think companies see that. That'll be the same kind of thing with AI systems; if you try to be proprietary and go it alone, you'll fall behind the rest of the open source. And so, it's much better to participate with the open source than try to compete against it.

The difference between AI and machine learning

02:25: AI is trying to write programs that do intelligent things. Machine learning is doing that by showing examples. And the alternative to that is an older technology we call "expert systems", which means you use the blood, sweat, and tears of graduate students to write down pieces of knowledge by hand rather than trying to learn them.

Data science is the intersection of statistics, machine learning and programming

03:00: I think of data science as a combination of statistics or machine learning, the ability to do some programming, but not necessarily be a professional-level programmer. And then expertise in the particular type of data you have, whether that's biology, economics, or whatever the data is. And so, data science is the combination or intersection of those three aspects.

Is there a possibility of generating revenue through subscriptions for big social media companies?

35:39: As a society, we still haven't really understood or adapted to how digital works. And people are super willing to say, “I'm going to spend $50 or even a hundred dollars per month for some kind of physical good that I pay to my phone or cable provider.” But when it comes to paying a few pennies to read something on the internet, it's, “oh, no. Information wants to be free.” And I think we might be better off in a world where these assets were all aggregated, and you just paid for a subscription.

Show Links:

Recommended Resources:

• Billy Beane
• Business Insider: The lawyer who used ChatGPT's fake legal cases in court said he was 'duped' by the AI, but a judge questioned how he didn't spot the 'legal gibberish'
• The New York Times: Google’s Photo App Still Can’t Find Gorillas. And Neither Can Apple’s 
• The New York Times: A Conversation With Bing’s Chatbot Left Me Deeply Unsettled
• robots.txt
• Massive open online course (MOOC)

Guest Profile:

• Faculty Profile at Stanford University
• Peter Norvig's Website
• Peter Norvig on LinkedIn
• ​​Peter Norvig on TEDTalk

His Work:

• Artificial Intelligence: A Modern Approach (4th Edition)
• Data Science in Context: Foundations, Challenges, Opportunities
• More scholarly articles

Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.