Lively Robot Teachers, Boring Human Teachers?

Carla Sarett, Ph.D.

I attended a high school that was, by most measurements, a "good" school – many of our students went on to Ivy League colleges and test scores were above the norm.  Yet looking back, I can recall only one teacher who made any impression on me whatsoever.  The rest of my teachers, whether in physics or Latin, might as well have faxed in their classes.  

It's a fact -- teachers who are engaging and stimulating are rare. Like acting or painting, teaching is an art, and when we've been lucky to have a great teacher or trainer, we've realized our good fortune. And so there is a constant public outcry that we must produce more wonderful teachers, and much hand wringing that great teachers are apparently in such short supply. 

Many solutions are proposed in order to "fix" the problem, including more tax dollars for education, higher teacher pay, privatization of schools, better teacher training, and better classroom facilities.  Add to this the seemingly endless debates about imposing national "standards" on teachers' classroom performance, to say nothing of the current mania for standardized tests.  And so it goes.  Parents pull their children away from public schools into private schools or home schools in the effort to "fix" the problem themselves.

But will these solutions really "fix" the problem?  At the most basic level, these solutions beg the question – what makes a great teacher great?  Certainly, teachers must be knowledgeable and competent (i.e., capable of transmitting information and managing a classroom), and the proposed solutions are targeted at cultivating knowledgeable and competent teachers.  But there is another equally important element to teaching:  the ability to engage and stimulate students.

When we think about great teachers as being not only competent, but also engaging, the goal of improving our schools seems even more daunting. Think back to high school or college or other training, and you’ll quickly flash back on how boring much of your experience was.  It seems unrealistic that we will ever have enough engaging, smart teachers to fill every course and every classroom.

Think about it: Even Corporate America, with all of its resources and inventiveness, has tremendous difficulty producing and maintaining an adequate supply of great managers.  This is true even though the salaries for executives in business are astronomically higher than teachers' salaries.  Companies devote enormous resources and training in order to mold a few individuals to be "good" managers.  Books get written on this topic, pundits regularly give new advice on how to identify and cultivate managers.  Yet it is no secret that talented managers are a rare commodity, and the many highly visible problems surrounding bad management are multiplying.   

This is not to say that corporate efforts toward improving management performance are wasted.  On the contrary, companies desperately need good managers more than ever – and probably need to devote more resources to producing them.  But the relative failure of corporate programs to cultivate management suggests that the paths to training future generations of talented human teachers are fraught with difficulties.

Of course there are many differences between training managers and educating teachers.  Preparing teachers, if anything, is more difficult. This is especially true of teachers beyond the early grades in which affection and supportiveness count far more anything else.  We may produce enough caring teachers but producing lively teachers is trickier.  More likely, as subject matter, especially in science and mathematics, gets more and more complex, we will have a shrinking supply of human teachers (at the high school and college levels) who are both trained and at the same time engaging.

I think it is far more productive for us to focus our resources on programming and creating lively and knowledgeable robot teachers and other intelligent interactive devices, rather than hope any particular school (or organization) will get its fair share of great human teachers. 

We'll never keep up with learning requirements if we only depend on humans. 

Paradoxical as it may seem, it is easier to make humans competent (i.e., knowledgeable) than it is to make them lively and engaging – and even more paradoxical, it will be easier to program machines to be personable and engaging than it will be train humans for these qualities. 

Why should this be so? 

On the face of it, it seems downright bizarre that it's easier to make a machine personable than to do the same for an actual human.  Aren't humans by definition more interesting than machines?  Wouldn't we all rather spend time communicating with other humans than interacting with a machine, however intelligent?

Well, not really.  For better or worse, many of us prefer reading listening to music, surfing the web and watching TV to interacting with friends and family.  Social critics lament this situation as a "problem" that demands study, even remediation, but it's more or less a fact of life.  Even when we are communicating through technologies like IM and cell phones, we are often doing other things (surfing the web or driving, etc.)    

Some humans are of course engaging – but the qualities that make certain humans engaging (e.g., humor) seem about as innate as any we can identify.  Botox can make us appear younger, and anti-depressants can make us less gloomy, but what can make us livelier and more interesting to others?  Creating more drugs to stimulate liveliness presents an unattractive vision of the future. 

Also, forcing (or using techniques to train) human teachers to be "lively" or "engaging" seems misguided.  Imagine the hurt feelings of informing young teachers-to-be that they are boring!  (And think how quickly the future labor pool would shrink as a consequence from this additional personality screening).  After all, there are many other qualities we demand of future teachers, such as dedication and competence in their subject matter

So, with humans, we have the problems of fragile egos as well as their certain resistance to personality-alteration.  Compared to the challenges of changing a real human personality, programming machines (robots) to have a personality is a snap.  Some of the tools we need for this are already in development at leading universities. In fact, many computerized experiences, like simulations, already have the high degree of involvement that learners want; and the online tutoring programs have the high degree of interactivity that learners need.
 
This is not to argue that machines can or should ever replace humans completely especially in K-12 education.  Far from it.  Only people can discipline other people, and only people can become emotionally involved with other people. Matters of discipline and affection will always belong to humans.   But discipline and affection, while they set the stage for learning, are not its center. A good learning environment requires that we engage minds and hold learners' attention, especially in high school and beyond.

We take it for granted that interacting with humans (like teachers) can be boring.  Again, think back to your memories of high school or college and the long lectures, the slide presentations, and the recitations of seemingly unrelated facts; even intriguing subjects were quickly transformed into an occasion for doodling and gossip.   Naturally, there were some group exercises inviting greater involvement but still you could not escape the lectures and lessons. 

Despite the fact that we know that humans can be (and often are) boring, we would never have a public debate about eliminating boring humans because well… we're human.  Nevertheless, boredom is not a laughing matter when it comes to learning.  It's a serious impediment to acquiring new skills and information.  Learning simply doesn’t happen when we are bored and inattentive.  

So here is where the debate becomes murky.  Parents may sense that their kids are bored in high school, and schools may realize that classrooms fail to engage learners.  At the same time, parents and educators alike are aware that computers, the Internet and other interactive media are highly stimulating and engaging – despite this, the resistance to replacing in-person learning experiences with machine learning is strong. 

The same parents who still groan at the boredom of their own high school education would become furious to discover that their kids attended schools where human teachers had been in part replaced by robots!  The administrators who perpetrated this scheme upon unwitting parents and students would surely be vilified.  Social critics, many of whom are now lambasting University of Phoenix and other cyber-schools, would be sure to pounce and lament on how students were being victimized. 

Because of this deep bias toward in-person learning, machine learning is often presented merely as a cost avoidance strategy, a sort of K-Mart alternative to in-person classrooms.  The deluxe route is in-person, but to save money, we're forced to make do with e-learning.  Deep down, we're sure we learn better (and more) from face to face interaction with people than we do from the most intelligent machines.  Presumably, talking, arguing and seeing people face to face teaches us things that, say, writing to people, does not.

This creates an awkward tension surrounding the use of technology for learning.  While everyone expects and demands that schools acquire computers and access to the Internet, they actually balk at the prospect of machines doing "too" much and students actually pursuing learning goals without teacher lessons and lectures.  Machines are supposed to provide a kind of super-library and set of databases, while human teachers do the "real" "in-person" teaching and classroom interaction provides the "real" benefits. 

Schools, by this logic, are expected to load up on expensive technology but teachers (and the magic of classroom interaction) are supposed to bear the brunt of instruction. Machines are the impersonal part of our education; humans are its "spark."  Our "gut" tells us that human-human interaction is what learning is about, even though most of us prefer reading books and surfing the web to talking to other people when we ourselves attempt to learn new things. 

In fact, the lively robots we create in the future may be better able to engage the minds of students than humans are.  They will be able to adapt to each learner's style and, equally important, their pace.  In mathematics, for example, learners may require purely visual explanations, surely more powerfully delivered through machines, as are branching as well as repetition.  It's even possible, as experiments by Clifford Nass and Byron Reeves (1996) suggest, that human learners will develop feelings of quasi-loyalty to learning machines.
 
Let's suppose we offered the average high school student the choice:  a super-smart lively robot or the nice average human teacher. Students could program robo-teachers to be funny, choose the human voice they liked, select a specific personality type (e.g., scholarly, funky, etc) and even choose their appearance.  Of course, robo-teachers could immediately understand questions and provide highly specific feedback.  Robo-teachers would adjust to each student's learning style and respond at the right pace. 

Or, students could elect to have a very nice but sort of boring adult teacher.   Which would the student choose? 

Which would you yourself choose if you actually wanted to learn a set of skills or a new important subject area?  Would you learn more interacting with other humans in a face-to-face environment than you would from an intelligently programmed machine?
 
Many of us cling to the idea that we always learn more and better from other people, and we need in-person human teachers to motivate and guide learning experiences.  But that's because we tend to think of intelligent machines (in the future, robots) as something divorced from humans.  We even become fearful hearing about friendly or lively machines -- it seems downright creepy to program robots with personalities. 

Our anxiety comes because we forget an essential fact: humans create machines.  Intelligent machines, our robo-teacher of the future, will acquire "human" kinds of personalities, voices and appearances largely because these devices are our own inventions.  As humans, we will inevitably extend the best parts of our personalities and minds onto teaching machines—and it is the best part of us that will create the next generation of learners.   Robot teachers will be able to interact with learners one on one -- adapting to the pace as well as the content interests of the learner.   

It's in our power to create this new learning world, and misguided prejudices should not slow us down.  With cheaper and more plentiful robo-teachers, the amount of individualized lively teaching that each student receives will increase by leaps and bounds.  Human teachers can shed those lesson plans and begin to interact, support and motivate students.  It’s a world in which both students and human teachers will be the winners.     

Author Background:
Carla Sarett is President of Internet Research Group, which she founded in 1997, and whose clients include New York Times, America Online, Dow Jones Interactive, Ziff Davis Media, Conde Nast Publishing, among others.  (www.internet-researchgroup.com).   She has over fifteen years of experience in marketing research and has a wide background in both quantitative and qualitative techniques.  Previously Dr. Sarett held research management positions at National Broadcasting Company and Home Box Office, where she directed research aimed at the development of original programming.  At Chilton Research Services, she initiated one of the first panels for web-based data collection studying the behavior and tastes of adolescents. 

Carla Sarett holds a doctorate from The Annenberg School for Communication at The University of Pennsylvania. She taught as Assistant Professor at The University of California at Santa Barbara and Queens College, The City University of New York, specializing and publishing in the area of research on children and television. 

Contact Information:

Dr. Carla Sarett
Internet Research Group LLC
351 East Conestoga Rd
Wayne PA  19087
Telephone:   610.971.2119
E-mail:  cjsarett@internet-researchgroup.com