so continuing our our discussion of what's going to happen with the future of work because of ai um so the the the issue here is that there are it's unlikely that entire jobs are going to be automated but certain percentages of those jobs will be automated and as time goes on uh jobs will change and there they'll they'll exploit human abilities more and more the ability to interact the ability to be creative and intuitive those capabilities will expand more and more and then what may happen over time is that is that a certain percentage of those jobs will be automatable through ai also so as i explained in the other in the previous lecture the way this this document works here is these bars represent the amount of these particular tasks that can currently be automated and then these numbers down here are the percentage of worker time spent in the united states doing that so only seven percent of worker time is spent doing something that is not terribly automatable but even so almost 10 percent of that is still nevertheless automatable and so i gave some numbers over here um ninety percent of what um people who do soldering and welding and putting in windows do for a living is automatable with current price with current technology 73 percent food service 47 percent of retail sales um bookkeepers accountants 86 of what they do can be automated what investment bankers do fifty percent of that is spent doing collecting and processing data and that's completely automatable um mortgage broker time ninety percent of it is spent processing applications well now with ml and with natural language processing that's going to be automated so what's going to happen with mortgage broker as a position well it may not exist okay so advances in natural language processing are going to continue to increase the proportion of automatable tasks so for example jobs in healthcare currently 43 percent automatable that's going to go up very soon to 66 percent and in retailing we're at 53 automatable now that's going to jump to 60 percent okay so natural language process is driving an increase in the proportion of things that can be automated well the good news um as i said jobs tend to go away but then other jobs tend to be created and so what are the jobs that ai is going to create well one of your readings breaks it down into what are called trainers explainers and sustainers so as we do put in these automations into place who is going to be responsible for making sure that they work and well not everybody is going to be a programmer because programming is going to be less and less required so we're going to need humans who understand business and who understand people to be able to do that training for example we looked at how robots can increasingly be trained by virtual reality where a person can simply go through the motions and ex you know explain through motion what the robot needs to do and that's how they'll train them okay so we don't need to have programmers we need to have people who understand processes so there are the trainers so the customer language tone and meaning trainer teaches ai systems to look beyond literal meanings detecting sarcasm okay the worldview trainer trains ais to develop global perspectives so that various cultural perspectives are continued okay so human abilities are going to be increasingly put into the machines and so we're going to need people who are able to do that and it isn't going to be programming so there are also explainers people who are going to be able to design decisions based on business context you know transparency analyst ai usefulness strategist okay um the um the web didn't exist when i was in college and the number of different jobs that have come up as a result of the web uh it's astonishing you know you can probably think of 20 different job titles that have to do with managing some kind of web presence that certainly did not exist and so these new positions it's certainly feasible that these will appear um finally sustainers automation ethicists automation economists evaluates the cost of poor machine performance machine relations managers promotes algorithms that perform well to greater scale okay so they're going to be a whole host of jobs that are going to involve training and clarifying and maintaining positions that are not necessarily technical positions but but involve being able to increasingly endow the machine with human-like characteristics