How to Innovate Your Career into Obsolescence: A.I.

Imagine an infinitely patient teacher, available 24-7, able to access vast informational resources at the blink of an eye, cognizant in the science of learning, and can adapt course materials to match your learning speed and style. Sounds like the dream of every parent. The learner might even find it fun because the coursework is designed to put the learner in the Zen state of “flow” even for the most challenging material. Sounds like the dream of every student. Personalized curricula will match your interests and skills, but embedded in your learning program will be the skills to think critically and creatively in a broad way as you adapt to a modern ever-changing cosmopolitan society. Sounds like the dream slogan of every politician. Or university president. Or technocrat. Or pundit.

This sounds like an impossible scenario, but our world might be inexorably moving in that direction. At least for those who have the good fortune of such an education. As we transition into an information-rich age with ever more powerful processing machines, the landscape of education is changing rapidly. Both my parents were schoolteachers and I don’t think the computer factored into any of their lesson plans. I started out as a whiteboard-only teacher (or blackboard, depending on the age of the classroom) with my lecture notes written out, and revised, by hand. I adapted quickly to the overhead projector (to show useful figures in black and white) and replaced it with PowerPoint slides in color. Textbook companies kept up with the changing times by providing transparencies and slides of their figures. The changes seemed gradual.

Learning Management Systems came along. I resisted them for a number of years, and now I use them in some of my classes. Then online homework bundled with the textbook helped ease the burden of grading, and allowed the students to get immediate feedback with a primitive system that could give hints or nudge them to correct an error. Now these systems (such as ALEKS and Knewton) are built towards adaptive learning, personalizing the lesson plan according to the learner’s base. Control of the curriculum has moved away from the instructor to the A.I. The questions and exercises have also increased in sophistication, as have the feedback systems. As the number of users increases, data analytics can be increasingly leveraged. Adjustments can be made in real-time and propagated in the blink of an eye. All these changes have come rapidly in the span of the last several years, and the technology is only getting better. (Here’s a link to an earlier post on technology trajectories in higher education.)

We are just starting to see the unbundling of higher education as online courses, coding boot camps, digital badges and certifications, from a variety of providers jostle for your time and money (promising you future time and money). What will be the fate of our “traditional” institutes of higher education? Will they adapt or die? If all we are offering is content (i.e. the one-way transmission one sees in a stereotypical traditional science course particularly at the introductory levels), this can be outsourced to an appropriate A.I. system, which might do a better job as a teacher than the human instructor. Even for classes that are discussion based, these will evolve to increase the role of the A.I. and data analytics, and decrease the role of the instructor. We teachers will be introduced to these as enhancements to our teaching, taking away some of the less enjoyable repetitive tasks, so we can “concentrate on the essential teaching tasks”. These “essential” tasks are ones that the A.I. isn’t so good at now, but will get better as it monitors how teachers manage discussions online. The day may come when we will increasingly become assembly line workers as the “best” curricula are consolidated through much pruning. Sure, there will probably remain a place for a small number of human designers to keep this up-to-date, but this will be a significant minority.

If you don’t think that A.I. systems will be sufficiently sophisticated, I highly recommend Jerry Kaplan’s book Humans Need Not Apply: A guide to wealth and work in the age of artificial intelligence. His involvement with the Stanford A.I. Lab and startups in Silicon Valley, that used A.I. methods in designing and developing new technologies, give him a bird’s eye view of the progress over the years. While the title of his book and some of the content seems bleak, his out look is provisional optimistic. He thinks we still have a window opportunity to “get things right” at this developmental stage particularly to address the looming spectre of income inequality that is likely to develop given the present course. Governments or perhaps agreements among large multi-national corporations might facilitate such measures.

Some highlights of the book (for me) were peeking behind the curtain of companies that make use of A.I. The power of A.I. was rather eye-opening in two areas: high frequency trading and digital marketing. Unbeknownst to me, there is a furious war raging in the digital world of bits and bytes where computer programs jostle for supremacy. Kaplan asks: “So what’s the root cause of all this electronic pandemonium – computer programs fighting each other over the opportunity to game our financial systems or influence our consumer behavior? Can’t synthetic intellects just play nice, like decent civilized people?” If I were to imagine this anthromorphically, it would be like a scene from the Matrix movies where programs do battle with each other in hand-to-hand combat and any weapons they can access.

Kaplan responds: “The answer is surprisingly simple. These systems are designed to achieve singular goals, without awareness of or concern for any side effects… there’s no incentive for combatants in these new electronic coliseums to show any mercy to each other, or to pay anything more than the bare minimum they must to get what they want.” He goes on to provide a series of examples, from what may seem benign to behaviors that we humans might find morally repugnant. The way A.I. learns adaptively does not follow the same process as humans, and it could optimize differently in a given situation.

Thanks to sci-fi movies I’ve thought that the rise of A.I. would be Terminator-like, and there would be an epic battle between humans and machines. Kaplan thinks this is unlikely and the “takeover”, if there is one, will be much more subtle. The “embodied” A.I., usually in some partly anthropomorphic robot entity, may resonate with us viewers in the entertainment world. However, this is not an efficient way of deploying an A.I. Kaplan argues that robots would work much better (and more cheaply) with a distributed network of sensors and tools. He gives a great example of a robotic housepainter.

“It’s easy to imagine a humanoid form climbing ladders and swinging a brush alongside its mortal coworkers. But it’s more likely to appear (for instance) as a squadron of flying drones, each outfitted with a spray nozzle and trailing a bag of paint. The drones maintain a precise distance from each other and the wood siding of your Colonial, instantly adjusting for wind gusts and other factors. As they individually run low on supplies, they fly over to a paint barrel to automatically refill and recharge, then return to the most useful position. A series of cameras sprinkled around the perimeter of the project continuously monitors this flying menagerie and assesses the quality of the job. The actual device directing this mechanical ballet needn’t even be present. It can be what’s called software-as-a-service (SAAS) rented by the manufacturer and running on the Amazon cloud. Why bother to put all that computing power out in the field where it may get rained on and be used only a few hours a week?”

Kaplan has plenty more examples, and if you like this excerpt of his writing style, I recommend you read his book. In the meantime, I’m wondering if I should think twice about delving further into the intersection of education and technology. While I’m excited about designing top-quality technology-enhanced curricula with adaptive capability to maximize learning, where the teacher becomes more of a guide-by-the-side coach, at some point I might be innovating myself out of a job. I knew I wanted to be a teacher when I was young, and honestly I thought that if I was good at it, I would have job security. We’ll always need teachers, won’t we? Actually I’ll probably be able to retire before such innovations completely change the playing field, but maybe I should think carefully about the human relational element and its distinct role in the learning process.