Welcome back to Frank & Sense! This week, I chatted with my buddy (and Google Research Scientist) Michael Madaio about all things AI.

But before that...

Frank & Sense is independently published and entirely supported by people like you. That means that your opinion matters to me! Could you spare a moment to share your thoughts via our audience feedback survey?

Take the survey now (all the cool kids are doing it!)

It takes only a few minutes and everyone who fills out the survey will be entered into a drawing to win $25! Enter now for your chance to win!

The deadline has been extended one final time to Friday, June 21.

Thank you for sharing your feedback, and I look forward to hearing from you! And now, on to the main event!

"Artificial Intelligence is a Metaphor"

Art by OpenArt.ai. Prompt: "It's a podcast episode of a show called "Frank & Sense." There are two people, Frank and his guest, Michael Madaio. Michael's cat Ruby is there as well. Cartoon style." I gave it a photo of Michael to reference, and this is what it spat out.

We've heard a lot about AI recently. But what is AI, really? This week, Frank is joined by his friend (and Google Research Scientist) Michael Madaio to discuss all things AI. Michael is a part of Google's Responsible AI group, and his research looks into how we can anticipate (and limit) risks and biases that may be perpetuated through AI. Michael gives a brief history of AI and shares some case studies of AI implementations across the country in various fields. I guarantee you will learn something by the end of this episode - I know I did.

Episode is embedded below and available everywhere you get your podcasts!

Stuff We Talk About:

We talk about so much stuff in this episode! I'm including all of the great articles that Michael references, as well as the resources that he recommends our listeners check out.

Articles

Machine Bias - This is the ProPublica article from 2016 that Michael describes as a "bombshell" in the AI world. This article shows how AI was being used to predict recidivism and how the algorithm was biased against Black people.
Inside the secret list of websites that make AI like ChatGPT sound smart - This is the Washington Post article that Michael references that shows the websites ChatGPT was trained on. As far as we know, ChatGPT was not trained on "Frank & Sense."
NYC’s government chatbot is lying about city laws and regulations - This Ars Technica article underscores Michael's frustration with the MyCity Chatbot from New York City. The chatbot was built ostensibly to help answer questions about local laws and regulations...but it knows about as much about city laws as I do. "Hey, I'm chattin' here!"