27/08/2019 by Rose Eveleth
Today we travel to a future where algorithms can be put on trial.
Guests: Rumman Chowdhury — Global Lead for Responsible AI at Accenture Applied Intelligence Kevin De Liban — attorney at Legal Aid or Arkansas Nicholson Price — assistant professor of law at University of Michigan Shobita Parthasarathy — professor of public policy and women’s studies at University of Michigan Actors: Evan Johnson as Mr. Morton David Romero as David Ash Greenberg as Ash Santos Flores as Santos Charlie Chalmers as Charlie Grace Nelligan as Grace Ava Ausman as Ava Sidney Perry-Thistle as Sidney Arthur Benjamin as Arthur
→ → → Further reading on today’s episode can be found here ← ← ←
Flash Forward is produced by me, Rose Eveleth. The intro music is by Asura and the outtro music is by Hussalonia. The episode art is by Matt Lubchansky. Special thanks to Veronica Simonetti and Erin Laetz at the Women’s Audio Mission, where all the intro scenes were recorded this season. Special thanks also to Evan Johnson who played Mr. Morton and also coordinated the actors of the Junior Acting Troupe who play the students in the intros this season.
Get in touch: Twitter // Facebook // Reddit // firstname.lastname@example.org
Support the show: Patreon // Donorbox
Subscribe: iTunes // Soundcloud // Spotify
Sponsor information: Skillshare is an online learning community with thousands of amazing classes covering dozens of creative and entrepreneurial skills. Flash Forward listeners get 2 months of unlimited access to thousands of classes for free by signing up at http://skillshare.com/flashforward Papers & books I mention in the Skillshare ad: “The art of being mentally healthy: a study to quantify the relationship between recreational arts engagement and mental well-being in the general population” and How to be a Happy Academic.
Learn more about your ad choices. Visit megaphone.fm/adchoices
Listen Date: early November
- Man, the debate club kids really did not get triage, optimisation, and resource constraints. It was frustrating to hear them talk.
- But I think the specific case the debate club was presented with wasn’t that great an example for the episode theme anyway.
- ‘Saved the family money in the end’ reminds me of the EconTalk episode with Adam Cifu on how medical treatments should consider benefits, side effects (or safety), and cost.
- ‘Put all our trust into machines’ – well, the other alternatives are to put it into people or into God, and have you seen people and God?
- I was at my aunt’s hospital’s 25th anniversary celebrations the other day, and the MCs were talking about how the final arbiter on a patient doing well or not, once doctors, medicines, and the patient were taken care of, was God. God must have really hated cancer patients back before the 1990s.
- The Arkansas algorithm sounds like the inevitable outcome of a lowest-bidder tender.
- Shobhita Parthasarthy’s opening statement about how we distrust ourselves, acts as a coutnerpoint to my earlier despair about trusting people. Heh.
- Actually, if someone was to now put skull measurements and IQ scores or Big 5 personality test scores through a neural net, what are the chances that they would come up with neophrenology?
- I should read up on the Cadmean vixen.
- I learned about two types of bias: in the first, systemic data is missing (the unbiased variable is still a proxy for a biased variable); and the original survey is itself biased. In the second, the bias is already present and the model is trained on biased data, similar to what the EconTalk episode on Weapons of Math Destruction talked about.
- I later discussed this episode with Jigar, and he expressed skepticism that any neural network is that much of a black box; because most NNs, he says, are first programmed around a decision tree.