Attended a couple talks at this event back on Tuesday. Wanted to share some quick notes and thoughts.
5:00 – 6:00 p.m. Keynote—Jacob Ward
This was by far the best talk methinks. @_jacobward_ started in behavioral science and took a journey over to modern events and emerging concerns. Ultimately he threw out a ton of great examples of human failings and how AI will be used by corporations to exploit and change us. It didn’t strike me as terrifying at the time, but my head was swimming for hours afterwards.
MORE DETAIL:
* Jacob starts by exploring the ways we all fail to make great decisions. Kahneman and Tversky researched intuition and “Prospect theory” (basically, how people are terrible with statistics) since the 50s. Kahneman went on to write this “Thinking fast and slow” book (just got a copy) after winning the 2002 nobel prize in economics (despite his being a psychologist, not an economist). He sums much of it up in this google talk (but it’s not a very engaging presentation. the book and the talk seem to quietly drop important terms and then get lost in examples, so it feels like he wandered all over the place. but he’s not. it just requires a lot of focus to follow)
* There are 3 important failings in human decisions under uncertainty (that we should all remember!):
1) Representativeness (we boil things/people down to a “type”)
2) Availability (we fall back on anecdotes and news exposure, instead of considering bigger picture facts)
3) Anchoring (car salesman trick: whoever mentions price first sets the conversation)
* Pol.is – a machine learning (yeah?) tool for politics. Sounds like it can scrape social media to deliver the top political issues from a community. (vTaiwan derives from it). Now used by reporters to choose the best hard hitting questions for politicians. AI in it’s best form: a speedy assistant.
* Joseph Weizenbaum – created the chatbot therapist (then was horrified when people loved it, and talked about not needing therapists anymore. Quit and became an environmentalist)(this guy needs a movie!)
*
“Superstition exemplifies how we capitulate with the things we don’t understand”
(!!!)(I’m dazzled by this from an HP Lovecraft perspective. And the baffling global power of antiquated organized religions)
* David Dao was pulled off that flight in 2017 (and pretty beat up). Apparently the computer that selected him ignored the legality of delaying a doctor with rounds in the morning. A security guard is now suing the airline for lack of training in how to follow or reject computer orders. (point being: we anthropomorphize computed results, and give them ridiculous authority) (* “the authority of the inscrutable” rises again!)
* Which version of your judgement do you want AI to believe? We are not always the same caliber of person. How is AI to determine which to empower and believe?
There are basically just 5 big corporations researching AI – and thanks to capitalism they would all rather empower the fast thinking (scared) version of you because that person is easier to monetize. (* I thought the military was also researching AI. was it william gibson who said we’re just waiting to see if AI will profit off us or kill us? because only corporations and the military are seriously evolving it? … i think it was someone else. but. i bet gibson is on board)
* Ultimately :
We are easy prey for the sorts of pattern recognition that AI rocks.
…uh, and I’ve discussed elements of the following two talks with coworkers so much that now I’m kind of exhausted to write out all the details again.
3:40 – 4:30 p.m. The future of AI and robotics: Opportunities and potential threats
* I though Joseph Orosco and Joshua Reeves had tons of excellent comments and challenging points of view. Finale Doshi-Velez got a couple good points in too, but Rebecca Hutchinson seemed quiet to the point that I was distracted wondering why she wasn’t chiming in more (not interested? irrelevant topics? personality? huh)
* Orosco pulled many great examples from pop culture and entertainment, but I think he was overlooking a big part of Frankenstein’s legacy. on another level the book is about the awful/horrifying things that would (or do) happen if men could create in the way women do.
* I disagree with Reeves’s comment that american’s won’t embrace self driving cars because we all think of cars as the embodiment of freedom. (on one hand i immediately think of the ridiculous numbers of deaths each year due to automobiles) (on the other hand i think AI will represent a new slave class, which is another arena where America also has notable history).
2:35 – 3:25 p.m. The tech economy: AI and robotics in the workplace
5pm
* thought this was the worst talk of the bunch. basically none of them seem that concerned (because while we’ll lose a bunch of jobs, we’ll gail a bunch of jobs).
* someone seemed to suggest this sort of employment shift was good for farmers. Which makes me think of recent news stories about the crisis of suicide rates across farmers worldwide. (though, this may be inflated? I tried to google up a link and instead found suggestion people were reporting this due to a 2015 CDC report that has since been modified. supposedly they meant to say that suicides for everyone in rural settings are up compared to suburban and city folk. not just farmers).
* in the end someone asked about universal basic income and I was slightly horrified that every panelist seemed against it. Jon Brewster claimed you cannot test it at a small scale before “universal” adoption (but hasn’t it been tested and proven in certain cities and countries in europe?). Alan Fern dropped the classic privileged capitalist bullshit that if your basic needs are taken care of you won’t have any incentive to get off your ass. so bizarre. (basically the complete opposite of how humans function. when you can’t take care of your basic needs, you fail to be productive).
(* ultimately, i think the whole idea that we give people money to spend how they want is the flaw. capitalism ruining everything. Deliver the food, housing, and services outside of the monetary system. Didn’t get the sense any panelist was on board with this crazy anti-capitalism perspective though. Doubt many Americans would be on board either. sigh).
…
anywho. might come back and jot more down later. just wanted to share these notes while the symposium is fresh in my mind. weeee