The current thinking about the automation of the workplace is that the jobs that require the most creativity will be those that are safest from the robot overlords when they come for our jobs. That might be true, but the erudite university professor with the rumpled corduroy jacket and scholarly spectacles might soon be headed for extinction as well.
The reason is that technology is already fundamentally changing the classroom and the way we deliver information to students, especially via the use of video. We also live in an era of the flipped classroom, where the role of instructors is morphing, as well as massive open online courses (MOOCs), where some university classes can have 1,000 or 10,000 or even 100,000 students in them at any time. What human professor can realistically teach these types of classes without some help from technology?
Michael Osborne, associate professor of machine learning at the University of Oxford and the co-director of the Oxford Martin Programme on Technology and Employment, says that robo-teachers may one day be the answer. As Osborne recently pointed out in an interview with The Telegraph, “It seems pointless to have a teacher or lecturer standing in front of a classroom statically delivering content that might be better absorbed through online videos, thereby leaving the teacher time to engage with students in a more interactive fashion.”
Moreover, robot professors might be able to personalize the information that students receive, as well as the types of problem sets or homework they must complete. That might keep the brightest kids engaged, while still making it possible to educate kids having the most problems picking up the material.
Delivering the information is the easy part, the hard part, says Osborne, is having the “deep understanding of human society” that’s required to be a teacher. Just think of that famous Jetsons episode from 1963, where there’s an impossibly complex equation written on a chalkboard, and the robo-teacher is staring straight ahead with a totally blank look. What machines can’t do right now, says Osborne, is provide the type of social and emotional intelligence needed to command an unruly classroom or respond to the non-verbal clues of students.
However, as artificial intelligence eventually catches up to the needs of robo-professors – something that even Osborne says conservatively could take as long as 20 years – expect to see professors as more of a mentor and facilitator, while much of the grunt work of lecturing and grading and test-making is done by machine. Students, too, will be called upon to do more of the functions currently done by professors.
At some point, it’s possible to imagine a scenario where the top, highest-paid human talent migrates to the nation’s most elite universities, while robots take over the teaching and training of students in community colleges or vocational schools, where most of the emphasis is on learning certain very specific skills that can be used over and over again in the workplace. The less creativity that’s required of students, the more likely is it that a robot could provide the type of instruction required to educate them.
But what does society really gain if super-smart robots result in the firing (or non-hiring) of academics on a productivity and cost basis?
On one hand, you could argue that society might be able to educate more students than ever before – and give them the kind of world-class education that is only available in a few institutions around the world. In a single MOOC, a top professor might teach more students than he or she would in an entire lifetime of teaching.
On the other hand, the educational classroom could basically turn into an educational factory, mass producing graduates instead of widgets. Instead of “teaching to the test,” educational institutions could start “teaching to the machine” – that is, focusing on the types of classes that are most easily taught by machines.
And if you buy into the whole robot overlord thesis (not so bizarre these days), then what if the robots start to decide what humans should study in order to be productive members of society? Robots might decide that learning low-level coding and programming skills is needed to pacify humans and keep them producing ever more machines. (Think Richard Dawkins and his idea of humans as survival machines for genes updated for an AI future). Human programmers coding the building blocks of artificially intelligent life could easily become the scenario for a dystopian Hollywood science fiction movie.
Probably one of the scariest articles written in 2014 was a piece in The Atlantic (“What Jobs Will the Robots Take?”), based on research from Osborne and his colleague Carl Benedikt Frey, who outlined the types of jobs that robots could take over in the workplace. Not content with hollowing out blue-collar and manufacturing jobs, robots could soon move into white-collar and other jobs in the creative industries. In fact, robots could take over nearly one half of all U.S. jobs in a decade or two. That scenario of massive workplace automation has recently been updated by Osborne and Frey, who suggest that 47 percent of U.S. jobs are at risk of being taken over by smart machines and software in the next two decades.
— Michael A Osborne (@maosbot) February 9, 2015
That has enormous implications for the educational system, which must educate students at a time of tremendous transformation of the workplace. Educators must race to prepare students for jobs that will actually exist in the future. It’s possible to imagine robots making it into the mainstream from the low end first, transitioning from tasks such as teaching kindergarten school kids in Tokyo and senior citizens in Dutch retirement homes, to eventually teaching kids in the classrooms of America’s universities.
A decade or two from now, a formal designation of tenure might be the only way to keep robots from taking over jobs in academia. Rather than being a form of prestige, tenure would really just evolve into a form of job protection and social guarantee. Publish or perish, indeed.