Monday, December 19, 2016

Teaching in the Machine Age (Or Not)

The Christensen Institute is devoted to "disruptive innovation," or as a five-year-old might put it, new and creative ways to kick over the big stack of blocks.

For the big stack of blocks that is public education, Christensen has the big boot of personalized computer-driven education-favored product. And a new part of their pitch is the recently released report/PR prospectus, "Teaching in the Machine Age: How Innovation Can Make Bad Teachers Good and Good Teachers Better" by Thomas Arnett.

Hi! I'm your new teacher, and here's my human assistant.

Arnett is a Senior Deep Education Thinker at Christensen, which is impressive since it was only five years ago that he was finishing out his two years with Teach For America in Kansas City. He then spent three months at Achievement First in 2012, and moved on to Senior Education Research Fellow at Christensen in 2013.If we dig deeper, we find that between 2002 and 2009 he was at Brigham Young earning a BS in Economics, and later did some graduate work at Carnegie Mellon's Tepper School of Business, including (and you'll want to remember this) Data Mining, Applications of Operations Research,Management of Software Development for Technology Executives, Innovation Ecosystems, plus Commercialization and Innovation: Strategy. So, no education.

This is the guy who's going to tell us how The Machines will help us do our jobs. Yay. But I have read this so you don't have to, because this trend toward Personalized Competency-Based Software-Centered Education has been poised to become the Next Big (Money-making) Thing, so we need to know what they're thinking. Here we go.

Introduction: Welcoming Our New Computer Overlords

Computers just go smart enough to win chess games and Jeopardy, so clearly they are ready to help us run the world. They are doing All The Things, which raises the question-- will join "assembly line workers, personal accountants, taxi drivers, sports journalists, and family-practice doctors" in the list of workers whose jobs "fall prey to machines." Wait! What? Family-practice doctors have been replaced by machines? I have got to pay closer attention to Dr. Fee the next time I stop in.

But fear not, teachers.Complex social skills are required for teaching, so we can't be automated. However, the part of teaching characterized by non-teachers as "dispense information and assess student's knowledge of rote facts and skills"-- well, Arnett hints he's not so sure you need a human for that.

But computer-centered education is still a solution in search of a problem, so Arnett proposes the problem-- getting a high-quality teacher for every student. He even proposes reasons for this problem. There are teacher shortages, fed by low pay and low prestige. Schools are bad at hiring the best teachers. Teachers are burning out and leaving what with the grueling hours etc etc, which exacerbates the shortage issue.

Fortunately, technology can help with these problems. Not by improving teacher pay, work conditions or prestige, silly. No, we're going to "commoditize professional expertise."


Commoditize. Which means basically the same as commodify, which means to turn into a commodity. Innovations can do this, and we're going to talk about doing it many, many times in this report.

These innovations simplify and automate some of the tasks of experts, making expert-quality work less scarce and more widely available. 

In other words, some skill sets are rare and therefor expensive and hard to get, like, say, getting Lebron James to play for your basketball team. If you could commoditize basketball skills, though, you could maybe come up with a hundred basketball-playing robots of comparable skill, and everyone could have Lebron on their team. In sectors like the restaurant biz, we solved this issue years ago by turning high-skill jobs (guy with skills and training to be a chef) into low-skills jobs (guy with ability to stick basket in deep fryer). But if we could commoditize teaching...

How could we do such a thing? What sorts of innovations would help commoditize teaching so that we wouldn't have to pay a bunch of money to highly trained professionals with a rare skill set?

Innovations That Simplify Professional Expertise

It used to be hard to diagnose scurvy. Now any mook can identify it by consulting with Dr. Google. Once we learn a lot about some deep, complex area of expertise, we no longer have to depend on highly trained skill sets to deal with the issues. Non-experts can just follow the rules and the procedures laid down by actual experts and-- voila!-- anybody can do the job just like a pro!

Innovations That Automate Professional Expertise

As the understanding of a field moves from expert intuition to rules-based practices, parallel developments in the field of computer science make it possible to automate many tasks that historically required the attention of experts.

After you've reduced expertise to simple procedures, any idiot can do the work like an expert. And "any idiot" always includes "anything run by a computer." Computers can now fill out your tax returns and figuring out your credit score.

What Happens To Commoditized Professions

Good question. Perhaps we should all ask the next gas station attendant we meet.

First, Arnett says that after commoditizationizing takes hold, non-experts can step in to do all sorts of work that previously required professionals. There may be a slight hole in his argument here:

For example, using rules-based medical science and the latest diagnostic equipment, middle-skilled professionals, such as nurses, can diagnose and treat many conditions

Yikes. I am absolutely not telling my mother- or sister-in-laws that as nurses they are just middle-skilled folks. If Arnett is ever in need of health care, I recommend that he keep this observation to himself as well.

Second, he says, the handling of so many tasks by computers and other drones frees up the trained expert to do more experty things. For instance, doctors can stop scanning test results and leave that to the computer while they go consult with patients about treatments and complications and other inquiries like "Holy hell are you telling me that my tumor diagnosis came from some machine and not an actual human!!??"

Arnett lays all this out in a chart to show that computers can do any algorithmic stuff, non-experts can just follow orders, and experts can use their "human cognitive flexibility" to handle creative problem solving and "engage in complex interpersonal communication."

Also, all the cool data and information that the computers and meat widget drone assistants collect will give the experts many more chances to come up with cool ideas as they sift through all that data.

See? There's still a place for carbon-based life forms in this brave new world. Just fewer of them, and cheaper ones. Feel better yet?

Arnett Puts His Foot In It

As he works his way around to explaining what all this has to do with teaching, Arnett says a Dumb Thing:

In industries, such as teaching, where professionals are under great pressure to do and accomplish more than they have in the past, assistance from non-experts and computers can be a huge boon to professionals

First, teaching is not an "industry." Teaching is "manufacturing" and it does not result in a "product," any more than ministers manufacture married people or couples manufacture children.

Second, teachers are not under pressure to do and accomplish more. Not really. We're under enormous pressure to waste a lot of time on malpractice like Common Core-based curriculum and tons of time doing test prep in order to get test scores up. So we are under pressure to do more pointless timewasting and less actual teaching.

Disney Animation

Arnett illustrates his point by talking about Disney animation, trying to show how the advent of computer tools gave animators all sorts of new capabilities and stuff they could do, shifting from the slide-rule computed camera tracking shots of Peter Pan to the CGI-guided swoops of Beauty and the Beast. He might have thrown in color issues as well-- the animators of Fantasia had to turn to things like fruit jelly to come up with colors they wanted, while modern animators have a rich and wide palatte to chose from. Hey, they've even made some big hits with this technology. I hear the young people really like the Frozen and the Finding Nemo.

On the one hand, I see Arnett's point. Technology has broadened the available possibilities for creators of animated movies, allowing experts in the field to imagine and create things far beyond what was available in the past.

On the other hand, his point is stupid because schools do not create commercial properties from scratch, but instead help develop and support live human beings. Because people aren't products. Animators start from a literal blank slate to create their products, and animated films don't have an opinion about what they want to be when they're created. Arnett needs to read that cute inspirational story about the blueberries.

"Create an awesomely marketable product" is not the gig in education. His analogy is terrible. But now he's ready to go back to the main questoin.

Will Innovations Replace Teachers?


If you were wondering what part of reformsterland Arnett hails from, he will now cite authorities like the Brookings Institute and Bill Gates.

Simplifying teacher expertise is no big deal-- heck, textbooks are an old tech version of that, saving us all from the trouble of coming up with our own materials. I actually have spent some time thinking about this, resulting in my decision a few years ago to stop using the grammar textbooks my school bought for our classes. I dumped them because they kept my teaching tied to their pace, their ideas, their examples, and their limited practice materials; I decided I would rather take my cue from my students and what they needed and how they could best be helped to understand. Could i have done this when I first started out? Probably not enough hours in the day-- and the fact that I can type materials up on a computer and have them printed out on a machine on the other side of the building certainly helps, so I guess I both object to and agree with Arnett's point.

Or maybe my point is that if you aren't very careful, labor-saving (or labor-transferring) technology will tell you how to do your job instead of helping you do it.

Arnett really wants to talk about automating part of the teaching process, having computers assign the materials and assess the materials and collect all the data from the materials, and not for the first time, I am kind of non-plussed by the way that folks like Arnett talk about all these hunks of adaptive teaching software as if it descended from a cloud, given some sort of divine breath by some higher power. I, on the other hand, view them as if some stranger knocked on my classroom door and said, "Hey, I'd like to come teach your class using my own materials." Is there any reason to believe that this stranger knows my job better than I do, that this stranger has any level of expertise that I should respect? I ask this question as an experienced teacher who has never yet seen a textbook that did not have its share of bonehead materials and just-plain-wrong baloney, clearly written and published by someone who had no idea how it would actually play out in a real classroom. Why should I believe that software will be better?

Software is written by people. Why should I trust those people or hand my classroom responsibilities over to them? There may be perfectly good answers to both of these questions, but we won't get to them if we keep pretending that computer software is magical and not just one more human-written teaching tool.

Arnett even has the nerve to bring up Pearson's WriteToLearn, yet another footnote in the long sad history of trying to get computer software to teach writing.

But now Arnett wants to talk about three specific situations in which it would be awesome to commoditize teachers and hook students up to a friendly computer.

#1 When Schools Lack Expert Teachers

Arnett believes that if you don't have a top-notch teacher to teach a class, some computer software would be great. He tells a long story from India to illustrate this point. He's heavy on how the human touch is still needed-- but boy do those magical computer programs make a difference, and I decide to take him even less seriously because he starts talking about measuring learning gains in some sort of linear fashion (these students learned had 2.5 times the gains of those students-- what does that even mean?? Test scores? Because I'm pretty sure India needs more than just people who do well on bubble tests.)

#2 When Expert Teachers Must Tackle an Array of Student Needs

Gosh, differentiation is just, you know, soooo hard! What would make it way easier is to have all of the students hooked up to computers that could automatically differentiate and spit out scads of data which the teacher can then pore through in her copious free time.

It's true-- meeting the needs of every individual student in the classroom is challenging. Oddly enough, Arnett doesn't discuss one obvious solution, which is to hire more teachers in order to create smaller class sizes, a known winner of an idea, but also an idea that doesn't make any money for all the people heavily invested in computer teaching software. More teachers and smaller class sizes would also provide no help to guys who went to Carnegie Mellon to study Data Mining (I told you to remember) which is so much easier when you've got the students doing all their work on the computer.

Arnett wants to sing the praises of Teach To One, one more teaching program in a box. I remain unimpressed.

#3 When Expert Teachers Need To Teach More Than Academics

So, always.

Increasingly, advocates are calling for schools to place greater emphasis on fostering students’ deeper learning and noncognitive skills. Recent research shows that noncognitive factors—such as goal setting, teamwork, emotional awareness, self-discipline, and grit—are strong predictors of how likely students are to persist through college and succeed in the workforce.

Oh, honey. This is the kind of thing that people say when they're young and impressed with themselves and they don't actually know very much at all about the teaching profession. This is the part where the TFA reformster comes breathlessly out of the classroom to declare, "Boy, those kids have like emotions and feelings and stuff and sometimes they have problems that they want to talk about and, man, I thought I would just teach math and reading and stuff but they really need other things and, dude, it's just hard!"

His exemplar here is Summit charters, a group that can barely acknowledge the non-academic needs of its staff, let alone its students. And Brainology. And project based learning and competency based education and basically every version of hooking up students to computers rather than humans. Or as I heard one parent comment, "I don't know whether to send the Christmas card to my child's teacher or to his computer." If the software (or rather, the software writer) is handling all the curriculum, practice and assessment, while the present live human is handling mentoring and non-cognitives, who is actually assisting whom?

Arnett offers data and theories. What he fails to offer is any evidence that computer-centered schooling is superior to any other model. In the end, he tries to dial it back and argue for a sort of teacher-cyborg, a expert (or near-expert) who is so computer-enhanced that she accomplish great new things with the help of her new technopal. Or maybe it's a computer teaching, and the human is just its attendant, and that's what is supposed to work so great. But evidence? There's none, just as there's no solid evidence that Arnett really understands what a teacher does in a classroom.

If you've been in teaching for more than a decade, you've been here before. Some salesman, who may have spent a year or two in classroom before deciding it wasn't really his gig, stops by to sell something. He doesn't seem to really understand your job, but he assures you that if you just buy his product and change your whole practice to match what the product is supposed to do, you will accomplish awesome things. Thanks, dude, but I'm already overstocked on snake oil.


  1. Arnett has a bit of a logistics problem on his hand that I don’t think he has come to grips with.Even IF his basic idea is workable (which it is not) – the devil lies in the following details:

    13 grades (K – 12)
    6 subjects per grade
    180 school days per grade

    That’s over 14,000 periods of code to write not including the nearly infinite number of options that qualify as “differentiation” in each subject. He also has a number of subjects that involve physical skills such as phys-ed, music, art, shop, lab sciences, etc. Any experienced teacher that has developed a proven program would want to go anywhere near his crap shoot.

    The fundamental problem without outsiders like Arnett meddling is that they have no idea just how complex a job it is to care for and educate hundreds of children, complete with their full range of humanity, day after day after day after day after day after day after day after day after day after day after. . .

  2. Peter, as you have noted numerous times before, this is all about who can get their hands on that sweet public money. Heaven forbid that it actually go to teachers!

    When we talk about exploding school budgets, much of the discussion focuses on legacy costs. However, I rarely see anything about the amount of money spent on tech upgrades. We have to pass bonds for 20 years of debt to address the technology that needs to be upgraded and replaced within 3 years. Obsolescence happens so fast and there's always a next great new thing.

    And school choice is the increasing rage, marketing (as you often note) will supplant school quality as the determinant of excellence. What better way to demonstrate excellence than state of the art technology?

    This also absolves administrators (and teachers) of critical decision-making. We can merely cite statistics as foolproof indicators. But the machine said...

    This isn't about quality. It's about trying to find some foolproof means of educating kids. And money, of course.