They're coming, and they'll be here by September! Robot teachers, programmed with a single mission: to save our failing schools.

Funded by the Frankenstein Foundation, computer engineers in secret mountain laboratories and workshops hidden deep below the desert floor are feverishly soldering chips and circuit boards onto bits of aluminum to create mechanical life forms whose sole purpose is to teach English.

We need this invasion of English-teaching robots because, according to researchers at the University of California, San Diego, "an unprecedented number of children in the US start public school with major deficits in basic academic skills, including vocabulary skills." So computer scientists at UCSD's Machine Perception Laboratory designed RUBI, a "sociable robot" who successfully taught a group of toddlers ten vocabulary words in only twelve days. RUBI improved the children's word-mastery by a full 25% compared to a control set of words not taught by the mechanical wonder.


In another experiment, another RUBI the Robot successfully taught English-speaking preschoolers nine words in Finnish (Finnish is notoriously difficult to learn because it is unrelated to any other language). And Korean educators report similar success with ENGKEY, yet another robot English teacher. English is mandatory in all Korean schools, and the government hopes to replace expensive and hard-to-recruit native speakers of American or Canadian English with expensive and hard-to-maintain machines like ENGKEY, a robot programmed to recognize and respond to human speech.

Above: The University of California's RUBI the Robot helps preschoolers build strong vocabularies in just 12 days! Below: The Korean robot ENGKEY listens and corrects


Like computers, robots appeal to school administrators who think the machines are smarter, cheaper, more efficient, and less likely to talk back or take sick days, than human teachers. And they appeal as well to legislators, government officials, and employers who are concerned with low test scores, high drop-out rates, and global economic competition.

To put the robot teacher invasion into context, we should remember that using technology to teach is hardly a new idea. Books are a teaching technology, though anyone who has studied a foreign language using only a textbook and then tried watching a foreign-language film knows that even tried-and-true book-learning has its limitations.

As for the newer communication technologies, when they came on the scene, telephones, radio, film, and television were all going to deliver information to students faster and more efficiently than any teacher could. What I learned from educational radio in the fourth grade was how to sleep in class with my eyes open, and what I learned from filmstrips in school was how long it took for the heat of the bulb to melt the celluloid when it jammed in the projector. Now it's the turn of computers and robots to fail at the task of educating children.

Learning languages with machines has been particularly attractive to those looking to spend less while at the same time reducing human interaction. In the 1960s consumers flocked to what was then considered a cutting-edge method, French-by-phonograph, or worse still, they played German-language reel-to-reel tapes as they slept, hoping to absorb the words subliminally. Even English speakers, scared by ads that warned, "People judge you by the words you use," tried mastering obscure English words by playing tapes on their automobile cassette decks while stuck in traffic. These have been replaced by newer but equally suspect machine alternatives to learning language from live human beings: computer programs like Rosetta Stone, language lessons on the iPod, and finally, robot teachers.

An 18th-century robot chess player known as "the Turk" is a reminder that robots don't always live up to the claims made about them: a human concealed under the chessboard was actually playing the game.

The Frankenstein Foundation has pumped millions into recombinant DNA research, but artificial humans have yet to hobble out of the labs it's subsidized. Even so, many observers are convinced that the foundation is right to spend its dollars to develop robot teachers. After all, children are naturally attracted to artificial life forms. Unfortunately, they don't always treat the robots gently. A singing, dancing version of UCSD's RUBI the Robot had its arms torn off by the over-eager two-year-olds on its first day in class. But when the robot was reprogrammed to cry if children came too close — like a car alarm — the children responded with a reassuring hug. Then they tore its arms off.

And the Korean ENGKEY robot places strong emphasis on the correct pronunciation of English, something that's often a problem for speakers of languages whose phonology is as different from English as Korean is. ENGKEY told one frustrated sixth-grader, "Not good this time! You need to focus more on your accent. Let's try again." If Korean schools proceed with the plan to stock their classrooms with robots like ENGKEY, entire generations of Koreans will grow up patterning their English on the vaguely central-European accents of the robot's machine-generated voice.

Just what one would expect from a machine-generated voice: ENGKEY tells an exasperated sixth-grader, "Not good this time!"

But while computers will continue to permeate every aspect of our lives and robots will continue building our cars, what may finally stop the robot invasion of our classrooms are new studies that suggest something that parents have always known: what children do most with computers is play games (much the same might be said of adults), and what they do with robots is take them apart to see who's inside. The New York Times reports that when children are given computers to help them catch up in school, their test scores actually go down, not up, because they're gaming, not poring over Latin or calculus. That doesn't bode well for initiatives like One Laptop per Child, which seeks to bring third-world children up to speed by giving them pre-programmed, indestructible laptops that can't load game software, or the many American programs that give low-income children home computers and internet access so they can match their more privileged peers.

Even if we learn most, and best, from one another, and by doing things, not lessons, we also learn a lot from machines, from telephones, radio, TV, movies, and computers,. So to play it safe, I think I'll take dance lessons from my Roomba, the robot vac that cleans my floors, just in case the much-feared robot teacher invasion really does come to pass.

Learn to rhumba while the Roomba cleans your floors. The robot vacuum has no arms or legs for curious children to tear off.

Dennis Baron is Professor of English and Linguistics at the University of Illinois. His book, A Better Pencil: Readers, Writers, and the Digital Revolution, looks at the evolution of communication technology, from pencils to pixels. This post originally appeared at OUPBlog and at Baron's own blog.