Can a computer lecture better than a human?

[cross-posted at the TechLearning blog]

I’m going to prime the pump a little bit for my K12 Online presentation next week…

My fifth-grade daughter’s math homework this weekend required her to find out what a radian or a grad was (hint: both are ways besides degrees to measure angles). We hit ye olde Google and quickly found this helpful (and free) learning activity from Wisconsin Online:

Bam! Ten minutes later my daughter and I had learned what a radian was (the animation was much more helpful than the mere definitions that we found), answered correctly all of the self-assessment questions, and were ready to move forward.

Okay, we actually weren’t ready to move forward because we thought the animation was so cool that we set up a Wisconsin Online account and dug around for other interesting tutorials. In other words, we were passionate, self-directed, engaged learners, prompted by a single question from my daughter’s math book.

I’ve thought about that moment quite a bit the last couple of days. Of course my mind started wandering to the K12 Online Conference, the TED videos, various podcasts, the MIT Open Courseware project, and other similar online multimedia resources. As my mind meandered around, it dug up a question:

Can a computer lecture better than a human?

That question’s not quite accurate, because there’s still a human behind every online learning activity. What I mean is that there are quite a number of examples on the Web of ways that we can learn and assess ourselves on fairly complex material using video lectures, animations, simulations, video games, and the like. As these resources grow in number – fueled by easy-to-learn, increasingly-powerful software that allows average citizens to create learning objects – and are organized and collected by individual experts, organizations like Wisconsin Online, or group efforts, it’s going to become unbelievably easy to find a variety of ways other than text to learn about almost anything we want. This will be especially true if we are intentional about it and actively work to fill in needed gaps.

Would I rather learn about a radian from a book or the Wisconsin Online animation? The animation – hands-down – due to its better explanatory power. Would I rather learn about a radian from a person or the animation? Well, the animation is infinitely patient – it doesn’t get irritated with me if I don’t understand the first time around. I can replay the animation as often as I need to but probably can’t ‘replay’ the person. The animation is more accessible – it was available to me in my home, at a time when I wanted to access it. And if the animation still doesn’t do the trick, there are other ways to learn the concept just a mouse click away (but other people who can explain radians usually aren’t so easily found).

Don’t get me wrong. There’s still a lot of value in human teachers when it comes to explaining difficult concepts, working through students’ misconceptions, inspiring students to want to explore deeper, and so on. We’re not replaceable by robots and software just yet. But it’s easy to see that simulations, animations, and other online text and multimedia resources can carry a great deal of the initial instructional delivery load.

There is a wealth of research showing that around 80 to 85 percent of classroom work is low-level factual and procedural work, exactly the kind of work that can easily be facilitated by the kinds of technology-mediated learning activities that I’ve alluded to this post. So why waste an expensive human on those things? If there’s going to be that much lecturing (and similar low-level learning work) going on, why not let the computer ‘lecture’ and free up our valuable humans for the stuff that software can’t do yet?

Comments

[M]

1.

I think the student would definitely have to be a "self-learner", but yes, definitely.

[Eloise]

2.

Surely it depends on what they're learning?

"What is a radian?" is fact based learning. It certainly has its place, but it much easier to develop computer-based resources for this than "Why use radians?" and other questions that indicate deeper learning or similar where a human to pose awkward questions, check out the concepts and correct misconceptions and so on is still king I think.

Can a computer teach certainly things better than a human? Yes, absolutely. Are we human teachers obsolete? Not yet - and maybe having good computer based learning resources to deal with many students will help the teachers focus on the less able and raise their standards whilst still challenging the more able ones.

[Wesley Fryer]

3.

Scott:

Rather than ask "Can a computer lecture better than a human?" I think a better question is "Can a blended learning context be more engaging, powerful, and effective to support cognitive growth and development than traditional, face-to-face instruction?" The answer is clearly a resounding yes. I am betting your interaction and conversations with your daughter during this learning process were very important to the final outcomes. The mere existence of this digital artifact was not sufficient to create the learning experience. Your guidance as a facilitator was also important.

Good questions and thoughts. Can't wait to share your K12Online08 preso with the world! Now I need to read Christensen's books too!

[sylvia martinez]

4.

The other commenters are right. The question is, why is it that ".. 80 to 85 percent of classroom work is low-level factual and procedural work."

That's the problem.

And it's not as simple as "she learned from the computer." You played a role, as did the conversation you had while you explored, and the interest you showed as you did this together.

By the way, if you and your daughter liked that site, you'll both love the Math Forum. Be sure to look at Ask Dr. Math and the Puzzles. http://mathforum.org/

[Scott McLeod]

5.

We are finding that large chunks of work that formerly needed to be done by people now can be done by software: filling out expense reports, registering new insurance accounts, evaluating mortgage applications, etc. Even high-end occupations that we think of as requiring cognitive complexity - like accountants (tax preparation software) and attorneys (legal document preparation software) - are finding to the dismay of their egos and their pocketbooks that substantial portions of their work CAN be routinized and ARE being replaced by software.

Teachers will NOT be immune from this. If a teacher is just going to stand up and lecture to students, a computer arguably can often do that better. If students are just going to sit at their desk, read a book chapter, and answer questions at the end (or fill out a worksheet), a computer arguably can do that better and/or make that more interesting and/or make that more efficient. And so on... I think we could collectively identify a number of aspects of teachers' jobs that are routine cognitive tasks that have the potential for being replaced by software (c.f. essay grading software or swipe cards / card readers for taking attendance). Why not have computers do those tasks and use our expensive humans for the complex things that software can't do? I think this is probably a good thing, because it frees up our teachers to do the important stuff.

FYI, as often is the case in the blogosphere, someone else has written about this in the past. Check out Hank Horkoff's old post:

http://snipurl.com/hank2007

[Joel]

6.

All questions about the 80%-85% factual recall aside (because I think THAT is the heart of the issue), I would ask:

When the animation realizes that you are not quite understanding the issue, what does it do differently for you?

(Sometimes we underestimate the human element and art behind what good teachers do.)

That being said, there is no question in my mind that multimedia type experiences can be more powerful than (what is the opposite word?) monomedia experiences.

[Jason Dyer]

7.

How is this different from, say, finding the material in a well-written book? Yes, there are differences, but -- the ability to bypass a teacher is hardly a new thing.

Incidentally, I can guarantee I can explain radians better than the video you just linked. Additionally can you answer the questions:

What are the drawbacks to the radian measure? What situations will you use a radian and what situations will you use a degree? What does the system of radians indicate that pi is in some sense "wrong" and the constant really should've been what we call 2 times pi?

Of course I could make a video telling you all that. But I would be telling; as a teacher I can lead you to get those answers above on your own.

[Scott McLeod]

8.

@Jason: I think the difference is that the computer can deliver the instruction in a variety of ways, not just text.

@Joel: If we build the learning activity right (not like the Wisconsin Online example), then it can adjust itself to the learner's success or lack thereof (like a video game does).

@Joel: Right. I think you should be doing the hard stuff. But the computer could do the initial instruction and get most of the kids to the place where you want them to be for the harder stuff. And instead of using your time for the intro stuff too, you could have done something else with your valuable time (like work with a few kids one-on-one). Or this could be their homework instead of using valuable class time...

[Gilbert]

9.

The computer offers multi-modal learning opportunities – so rather than having the need for an adaptive teacher, the students can just shift through the learning style they prefer. As well as ‘rapidly prototyping’ solutions and getting instant feedback – and if written like a great video game then the system will teach in motivationally sized sections set by the learners ability.With much more positive reinforcement than the average teacher of 30+ students could provide.

It is simple: ‘Any teacher who is replaced by a computer deserves to be!’

My question is are we really seeing ‘education’ through these isolated (as opposed to digital extensions of the analogue classroom) online systems or is it just ‘trainning’?

Education takes a lot of 'human'.

[Jason Dyer]

10.

I realize the book *medium* is different, but the conceptual idea of learning from a different source from a teacher is the same.

Dynamic adjustment, on the other hand ...

If we build the learning activity right (not like the Wisconsin Online example), then it can adjust itself to the learner's success or lack thereof (like a video game does).

... is indeed different from a book. That's the PLATO software, essentially.

From my experience it is great in some cases but it does have limits.

[Marshall]

11.

@Scott
Thanks for the intellectual work out. What is the mental equivalent of Gatorade? Knowing your strong bias to increasing tech in schools, I believe that the phrasing can be overlooked. Base question of "better" is riddled with caveats that have mostly been explored already. Bottom line, using tech is still a tool for good educators, so are text books, maps on the wall, and (my favorite) personal interaction. Although I agree that pushing/encouraging technology in the classroom enhances learning, the concept of removing the personal aspect is a legitimate concern of the over-application of this effort. "Intelligent, socially inept" adults is not what our goal should be either.

[Ken Allan]

12.

Kia ora Scott!

I didn't comment on this post when I first read it, for I knew where it was coming from. With computer instruction, the adage, "a place for everything and everything in its place" fits well.

With (digital) learning resources I bear in mind Death by Chocolate. A poor chef includes chocolate as an ingredient in every dish but it is a shortsighted one who excludes its use altogether. If the only recipe available that includes it is a mediocre one then chocolate should be off the menu. A good chef chooses recipes wisely.

Ka kite
from Middle-earth

[Jeff]

13.

This is something that I have also been curious about as of late. The computer offers so many things that text or a teacher cannot. But, the power of human teaching cannot be denied as well. I have found that the do it yourself approach has not worked for me, or the students I have worked with. I believe that computers are best as just aids, but should not be the primary source for education.

[Frank]

14.

Scott, I fully agree. I am taking a bit of different perspective on this though - regarding your point "So why waste an expensive human on those things?" – I would add, in many, many schools around the world there is no suitably qualified teaching staff. I have met some "English teachers" in rural India, for example, who barely spoke English. I think it is in such schools where computer or video-based teaching has the most potential, not to replace the teacher, but to complement the teacher's knowledge and abilities. This was one of the main motivations for setting up EduTube (http://www.edutube.org) - to organize some of the high quality educational video content which is already out there, so that it could potentially be used in such settings. Of course, in areas where there is a lack of funding for teacher training, teacher salaries, school resources etc., where is the funding for computers going to come from? I think that despite recent developments, the technology is indeed not ready yet for usage in such areas (in terms of costs / usability). My own experiences have led me to become much more pessimistic about the possibilities. However, this is also why I became more interested in video - because it is more low-tech, cheaper, and fairly straight forward to set up and use, and can be shared by many rather than used by individuals.

[Caleb]

15.

I enjoyed reading this article. You provide some interesting information about how technology can be an extremely useful learning tool.

[Kim W]

16.

Just a reminder that, in addition to that "expensive human" teacher, there is likely to be at least one and probably several additional "expensive humans" required to create these multimedia experiences! (And I don't at all underestimate the potential value of the animation Scott and his daughter viewed.) I think the important point is that instructors don't need to present every bit of content in class; whether in a textbook, web site, animation, etc., a lot of well-designed, well-developed content is accessible to students outside of class. Then, having gotten the basics (I'm thinking here of the lower levels of Bloom's taxonomy) out of the way, that valuable face-to-face time in class can be used more productively to provide help where needed and to help students work through some of the more complex issues that many have brought up here, e.g., when and why would you use radians vs degrees. In addition, by expecting students to "cover" some of the content outside of class and then providing in-class activities that utilize and apply that content, you help students learn to take responsibility for their own learning.

[Cheryl Oakes]

17.

Scott, Thanks for beginning the conversation and more thanks to all the comments! I am so glad that you provided the link where the conversation can continue.
Recently, my district just purchased an online typing program. (never mind that we are still teaching typing). The student logs in, at home or school, chooses their avatar, chooses their theme, assignment, etc. The human teacher monitors lessons, wpm, direction, and can make some adjustments. However, the student is in charge of the choice, the buy in and the ultimate learning.
Our class time provides for the initial setup and introduction. The student ubiquitous access allows for continued growth. Blended options, access and opportunity will all allow for facilitated learning of many subjects, not just the subjects that a school decides to "teach".

[TRM]

18.

Interesting that none of the viewers of this pleasant little graphic actually learned enough to notice that one of the slides in the presentation was incorrect.

Interesting, too, that wisconsin-online has not corrected the mistake.

The real problem is that this is touted as the paragon of education. "Bam! Ten minutes later my daughter and I had learned what a radian was (the animation was much more helpful than the mere definitions that we found), answered correctly all of the self-assessment questions, and were ready to move forward."
or
"It’s going to become unbelievably easy to find a variety of ways other than text to learn about almost anything we want."

Well, not really. You found a definition and answered three multiple-choice questions.

Let's look in more detail at the "wonderful job" this thing does.

page 3 says "click next to learn why they chose 360" - and shows a circle divided into 72 sections. Page 4 puts out a reason that is NOT the most likely and doesn't mention the others. These others are more likely and more reasonable if you credit the Babylonians with any brains at all. Unfortunately this animation just assumes they miscounted the number of days in a year. Interesting.

The animations ARE useful. I would probably use one myself if I hadn't already had several versions of it. Mine don't have the straight red line extending a bit as it gets wraps around the circle but some kids wouldn't notice that happening.

I also don't like the habit of filling in the sector to represent the angle. It's an unnecessary multimedia "feature" that doesn't help understanding. That's probably just quibbling, though.

page 12 is really quick with no additional explanations - I can't see students accepting that fact without questions or discussion.

page 16 is just wrong. Think about it -- page 14 just got done saying that pi radians is a half-circle. (That's 3.14 radians for those asking) The page 16 figure has three radii as being MORE than a half-circle. This is such a simple error to fix, yet no one has. WHY? Can it be that the creator did not realize the error himself and that no one has pointed it out? Again, why has no one pointed it out? - because all the information on the internet is true, as far as most students are concerned. Who dares question the great shiny box?

This is the problem with computerized learning, internet information, wikipedia, and many other "modern" methods of learning -- it seems so good because it's colorful and mostly accurate, yet contains mistakes that are fundamental and should have been obvious. These mistakes arise because the programmers are not math people, and math people are generally not programmers. More, those programmers are very often teenagers with an imperfect grasp of the material, but have the time to waste creating Wikipedia content.

It seemed somewhat amateurish, like something I'd get if I assigned it's creation to my 11th graders. The fact that it's held up as the model for the next generation of teachers is disturbing and bears witness to the folly of depending on the top results from Google to answer much more than trivia questions.

Fortunately for the daughter, trivia was all that was needed. She is in 5th grade, after all, and her teacher wouldn't have noticed this error.

[Scott McLeod]

19.

Thanks for the thoughtful comment, TRM. Of course textbooks have errors too, and often teachers miss those as well. We can't blame the medium (Web or print), only our authors and fact-checkers. I appreciate you giving me some food for thought for today!

[TRM]

20.

As quick reply and follow on, then I'm going back to checking homework papers.

Fixed stays fixed.
I agree that books have errors, but mostly in the problem solutions or in some area that is left to the graduate students to flesh out. The teachers I know take great pride in finding those errors and the students great satisfaction at noticing and correcting an error in the supposedly perfect text. An error, once found, remains corrected. Internet errors, especially Wikipedia errors, have an amazing ability to resurrect themselves. I am reminded of the case of a man who couldn't correct his own biography - the teenaged uber-user kept rewriting it and finally locked him out. How many other errors are there?

Rarity
The second issue with "Textbooks are wrong, too" is the frequency. True, textbooks have errors, but far fewer than the information I typically find on the web.

Intent
Book errors are usually due to oversight or grad student miscalculation. They are rarely errors in fact or method. The errors on the web are due to not only typographical or "understandable" mistakes - I would include page 16 above - but also of the intentionally misleading (what politics calls spin), intentionally incorrect and nasty (see martinlutherking.org) or just plain boneheaded nonsense.

I don't mean to pick on your one line, but the flexible and ever-changing nature of the medium demands that factual and methodological errors be fixed as soon as possible. I don't see that happening in very many cases.

Does internet reference ever work? Wolfram, yes. Wikipedia, only when it's a fact-based question that can be easily checked. Wisc-online - as far as I can tell. General websites, almost never - you just can't trust the answers unless you already know them.

Internet learning can be valuable - fix the errors above and that would be a good explanation of radians. But how are students supposed to know what's good and not? The previous commenters and reviewers and website users all took that little tutorial and none realized the errors. And that's a single page on a generally good site.

These ARE the fault of the medium as it is today and we must lay the blame at its feet or the medium will never change or improve. Is it a deal killer? No, but it worries me.

The whole lipstick on a pig thing.

But then, I don't want to be political.

[Pat]

21.

Great post! After reading some other posts, I wondered if it wouldn't be more effective for kids to listen to teacher lectures at home for homework and during class engage them in discussions and activities that go with the lecture. Of course this would involve more teacher planning and work, but isn't that our job?

[Andrew Brown]

22.

I'm a self learner and always has been. I wish teachers played the role of coach more than instructor while I was in school.