Chalkface Blog

John Battelle draws my attention to W. Daniel Hillis’ 2000 essay “Aristotle”. In it, he postulates an automated online tutoring system (“Aristotle”) that would find available resources and, on the fly, construct a completely individualised course for the student on the topic they wished to learn.

It’s an extraordinarily bold vision; one that contains enough challenges that no teacher need start worrying about job security just yet.

Arostotle’s constructors would face two challenges; course design and student assessment.

Course design

Here’s how Hillis envisions it.

Aristotle plans its lessons by finding chains of explanations that connect the concepts you need to learn to what you already know. It chooses the explanatory paths that match your favorite style of learning, including enough side paths, interesting examples, and related curiosities to match your level of interest.

Aristotle must know what explanatory material is out there. An experienced Googler would find this quite plausible; so far so good. Next, it must be able to rank this material for its appropriateness to the particular student. That’s harder, but again the most advanced search engines, such as Amazon’s A9, are making serous attempts to personalise the search process to this degree.

Finally, it must present these resources in a logical chain. Hollis’ idea for this

Whenever possible, Aristotle follows the paths laid down by great teachers in the knowledge web.

doesn’t work in my opinion. I think it takes us right back to where we are now; precreated courses such as Chalkface’s own online GCSEs. One might get around this, though, by asking the student to sequence a list of summaries by hand. Ultimately, the course design stage is do-able, perhaps not yet, but in a few years’ time.

Student assessment

Aristotle will not only explain things to you but will also ask you questions—both to make you think and to verify for itself that concepts are being learned successfully. When an explanation doesn’t work, Aristotle tries another approach, and of course you can always ask questions, request examples, and give Aristotle explicit feedback on how it’s doing. Aristotle then uses all these forms of feedback to adjust the lesson, and in the process it learns more about you.

Here’s where the hard work really needs to be done. Think back to your Bloom’s Taxonomy basics. In assessing a student, you are looking for three things

• Knowledge
• Understanding
• Evaluation

Simple questions will only test knowledge. To go further we need a deeper process. Our own work here at Chalkface on essay marking has shown that understanding and evaluation can also be tested for, using applied psycholinguistics, but only under closely controlled conditions.

I predict that this is the factor which will limit Aristotle’s development. A sufficiently general assessment engine is at least 10 years away; probably more.

Hollis’ own inspiration comes from Neil Stephenson’s novel The Diamond Age. Stephenson’s automated teacher

becomes a friend and playmate to the heroine of the novel, and guides not only her intellectual but also her emotional development.

To achieve this, we’ll need machines that are in every way as intelligent as humans. I don’t care to predict when that will be, but for an exploration of how it might come to be, I refer you to William Gibson’s classic The Idoru, which I’ve just finished rereading for the nth time.

Back in November 2000 I told educational publishers’ conference that the writing was on the screen for the paper textbook. From the conference programme:

Ian Grove-Stephensen, Managing Director of the Chalkface Project, argues that any paper-based investment made today will fail to earn a return; but reveals how publishers can leverage their existing resources to make real web-based profits now.

My reasoning wasn’t particularly esoteric. I simply noted that Moore’s Law was going to make computers cheaper than books. Nor did I think it was that big a deal for publishers. What we publishers do for a living is package up information to make it more useful to different kinds of users. I didn’t think then (and don’t think now) that the new medium was going to change that basic principle.

The audience unanimously disagreed and my presentation bombed sufficiently badly that the organisers didn’t ask me back. Most of them didn’t see it happening in their lifetimes, let alone in the five-year timescale I laid out. I left the conference very firmly put in place by my superiors.

Thus it was with guilty schadenfreude that I read this article in USA Today. American schools, they say, are about to reach the tipping point where computers are cheaper than books.

“A child’s set of textbooks costs $350,” Smith said. “If they can get these notebooks down to $500, it gets cost-effective in a hurry.”

The article doesn’t make it clear over what timescale that $350 is spent, so we have to be a bit careful translating it for the UK. Here we spend very approximately £20 per student per year (PA calculation). Notebook computers are still expensive, but desktops are not. I’ve argued elsewhere that they are a better bet anyway. The price of a desktop computer is down around the £200 mark now. So a computer would have to last 10 years, and the information on it would have to be free, for us to be at the tipping point now.

But by start of the academic year of 2005, that computer will be down to £100 (Moore’s law again). Which is only a five-year payback. And whilst there’s still the information to pay for, against that the computer can provide access to far more of it. So my prediction is that the tipping point will come in September 2005. From then on change will accelerate as schools grasp the economic opportunity.

A couple of years ago, I asked Berthold Weidmann, esteemed director of NETLinc, whether he thought school students in Lincolnshire would all be carrying laptops by 2007. This was a target being promoted by Microsoft’s Anytime Anywhere Learning initiative.

His answer, unequivocally, “No.”

“But” he continued, “it doesn’t matter because they all have mobile phones.”

I’ve been puzzling since over just how Berthold’s vision was going to develop. It’s not the devices I’m worried about. There are several PDA + Phone combos on the market now, and by next year prices will have dropped to levels where kids will buy them. What’s concerning me is the cost of bandwidth. I don’t believe either schools or parents will be willing to pay even a small fraction of the costs that online lessons would cost over a GPRS or G3 network.

Now, reported in the Korea Times, (via Anthony Townsend), we can see a model starting to form. The Korean government is subsidising its phone operators to offer this gadget at a knockdown price. The significance is that this phone/PDA will access Nespot in preference to the phone network if it can get a signal.

Nespot is a national network of Wi-Fi hotspots. Korea Telecom (KT) say

At present 12,000 NESPOT zones (Hot spots) use NESPOT service and wireless high-speed Internet service.
Other areas use [mobile phone] service, so that users can economically use the Internet anywhere, any time, in Korea. This is the actual service launched last February by KT and KTF.

So, the Korean model is

• a national network of low-cost Wi-Fi hotspots.
• a usable browsing device that can switch between phone (expensive) and Wi-Fi (cheap) networks.
• a government subsidy sufficient to make the whole system popular, which will bring costs down further.

It takes very little imagination to fill in the last bits of Berthold’s jigsaw puzzle. Put Nespot Wi-Fi hubs into every home (in Korea they cost $20/month which is about half the price of an ordinary ADSL connection here) and into every school. They are already in public transport nexi and other places where kids congregate. Inspire the content creators (e.g. me and my editorial staff) to produce stunning materials that really use the networked abilities, and which compensate for that small screen.

And you’ve done it. Anytime, anywhere learning achieved with three years to spare. Provided you’re in Korea, that is.

Saqib Akram, a very proactive teacher from Derby High School who uses our online GCSE Applied Business course, emailed us today with some feedback from his students

The wording on all assignment questions eg. Module 1 portfolio A Cameron balloons. All pupils felt that the questions asked could be simplified as they found it difficult to understand. They also felt that question had too much text in them which put them off reading all of the question.

We’ve also had equal but opposite feedback from other teachers who fear that the work does not stretch their students enough to aim for an A*. All entirely normal stuff; students of that course cover the whole ability range and you can’t please everyone. Or can you?

Right from its inception, Paperless School was designed to invisibly manage differentiated material, applying the right level of a course to each student, without involving the teacher in a load of dreary admin. Challenge number one, which with Saqib’s impetus we’ll address over the Summer, is to produce parallel versions of the core course modules.

Challenge number two is more vexing. How do we know each student’s level? To begin with, it’s easy. We simply ask the teacher to predict that student’s grade. But what if the student improves? Or backslides?

We could use each student’s current marks to determine a level, perhaps. But then one fluke good grade could bunk the student up into a level where they flounder, or vice-versa. Perhaps an average over the last three assignments would be enough to demonstrate a level that was both current and sustainable. We’re going to have to experiment widely with this before we get it right, I think. Until then, it will still be down to Saqib and his colleagues to periodically update the system’s knowledge of his students’ abilities.

NLP Conference

My introduction to educational publishing came when I co-authored a book of worksheets on self-motivation and related topics, called Goal Setting and Decision Making. I found the process sufficiently rewarding that I acquired the publishing rights to it, along with some 39 other titles, and used them to found the Chalkface Project. I’m proud to say that it still sells well 15 years later.

That book was based on a (then) very new branch of psychology called Neuro-Linguistic Programming (NLP). I had recently completed training in this and it had made a huge difference to my life. I wrote the book because I wanted to find a way to give back.

I still feel that sense of obligation, so when I was asked to develop the website for November’s NLP Conference in London, I naturally agreed. That website is finally ready, and you might be interested to know that it features a very interesting-looking track on NLP in Education.