What if we let AI do the thinking?

what kind of task does AI automate then if we see AI as this whole progression to augmented individuals if you will to individual creativity and productivity which tasks are on The Cutting Board for this I feel like I am bound to introduce a new word because we don't have good words for this we we think about whole jobs that's the wrong way to think about it see for every task there is a thinking piece and what I like to call a thunking piece and thunk being like the dull thunk of a brick hitting the ground that mindless brainless task or think of it as you want to go through the door you could you know use your head as a battering ram and thunk thunk thunk on the task or you could do a little bit of thinking and see what you're dealing with and realize that maybe the door handle is the solution to the problem that's the thinking bit and then once you've thought and you've seen the solution yourself and to actually execute and pull on the door there is more thunking again AI does not automate thinking it doesn't there is a lot of strange Rumblings about this that sound very odd to me who's been in the space for two decades but it doesn't automate thinking it does automate the thunking component of a task and you might be like well hang on Cassie hang on hang on what about art is that not are we not automating creative stuff there look if you think that we are replacing the thinking that artists do I think you've missed like a whole Century of art history so let's go and look at a quick minute of art history for you so this piece is a very famous piece called Fountain 1917 by Marshall duchenne this is in iconic art piece there was a big thinking component here which is why this is such an icon where the artists had to think about what needs to be said in the artistic conversation what is important what are the ideas how do I capture this and convey it to others and then he could have actually gone and learned how to work with porcelain and he could have carved and sculpted this himself but he didn't he took a ready-made urinal he signed a name on it not even his name and brought it to the exhibition so who was the artist there are we going to say that it is the toilet manufacturer that is the artist I would argue strongly against the thinking component was knowing what is worth saying knowing what is worth doing we are thinking when we are coming up with ideas when we are solving problems when we're being creative when we're interacting with one another not in a we pre-planned our speech and we're not we've sort of checked out of the conversation but when we're having true conversations with one another we're focusing we're reacting that's when we're thinking when we're inspiring and being inspired all of that remains and I believe that that is the best of how Humanity can spend its time and it is often when we are thinking that we're distracted from our opportunities to think to solve the world's most important problems that anyone who says Hey there'll be nothing left to do after AI after automation after productivity Revolution hasn't seen just how much is sub-optimal I'll call it sub-optimal in the world as long as we have medicines that have been undiscovered diseases uncured inefficiencies we have as long as we have climate problems as long as we have a future to invent there is plenty to think about and plenty to do we're gonna have plenty to work on and there will always be stuff to do there will always be ways to participate but also secretly this is rumbling about this concept of secret cyborgs I don't know if you've heard this this concept is essentially that people would love to get their thinking automated once you see the solution you would love to say okay now someone else do it now ai do it now machine do it you'd love to do that but you're scared that your boss will find out that might be the biggest problem that management has not learned how to manage thinking only how to manage thinking what you can measure easily which you can observe easily how many hours has someone sat in a chair how many times have they done a repetitive task whereas thinking is something you can't force you can only get in the way of it I don't believe that there is a person who can sit and think truly at your full creative potential for nine hours a day you need that space in between we should be okay with that we should be stepping up as managers and as teachers too and thinking about how to lead in a society where there might be some space around the thinking and that's okay now you might be saying okay but what if we let AI do the thinking on philosophical grounds I completely reject this notion AI can not think a and I can automate and AI can execute decisions when you are not the one making the decision and it looks like the machine is doing it there is someone who is actually making that decision for you whoever built the technology that is the person doing the thinking and I think that we've been complacent we've allowed our technology to be faceless even if I have a a cup that someone designed stupidly has a hole in the bottom I pour coffee in there it goes all over me if I react to this I'm like ah stupid cup it's a dangerous reaction the better reaction is who designed this thing this way let's have a conversation about that is that a right a useful design for users for society now with a cup it's bad enough but at least it's easy to do to go from the object to the person but what about these large-scale systems well let me tell you there are definitely people who are making subjective decisions designing these things we need to start seeing them we need to stop allowing technology to be faceless when we program the traditional way how do we get the task done some human being has to agonize over every single line so you want to automate a task it takes 10 000 lines of code 100 000 lines of code either you have to write them all and think about them all or some other human being has to have done that and you take the package and plug it in but when it comes to AI there are actually only two lines behind all that fiddling optimize this goal on that data set go two lines of thought that can scale up and affect millions of lives billions of lives these two lines require as much thought and respect as those ten thousand or a hundred thousand lines of code and if we don't see the people who are doing this if we say ah the machine makes the decisions it's a bunch of mathematics of course it's objectively correct how will we hold them accountable for quality for wisdom these two lines are fundamentally subjective what is success how do we make trade-offs what's good what's worth doing what isn't which mistakes do we tolerate and which mistakes do we not very subjective decisions right there that require a lot of decision skill and on that data set which date is it which examples do we want to teach with which examples are inappropriate what kind of world are we describing to the machine and is that the world that we want to build later this is a way to amplify the impact of decision makers so let AI stand for Amplified impact of individuals of users of everyone who's going to use it to be more productive but also of the people for whom this is a grand lever I hope that we will step up and understand that thinking is our responsibility and the people who design these systems I hope that we will hold them accountable for bringing the best of humanity to that design rather than the worst of us so I'll leave you with that thought this is about thinking and we will all have space to think more in the future it's time that we took our skills at it seriously learn to become better designers better decision makers and let's step up to it and build a better future together thank you very much [Applause]

Channel: Cassie Kozyrkov

Views: 4,735

Keywords: DataScience, Data, DecisionIntelligence, MachineLearning, Statistics, AI, Analytics, Google, GoogleCloud, Education, ArtificialIntelligence, Decisions, Leadership, Technology, Cloud, Cassie, Cassie Kozyrkov, Tech, Google Cloud, Cassie Google, Google Cassie, AI Cassie, Cassie AI, Cassie ML, Data Cassie, Cassie Data, Data Science

Id: OZlfDlZ1oSw

Length: 9min 40sec (580 seconds)

Published: Sat Jul 29 2023