A broken system

Today’s post is about how technology might disrupt the education system.

  • We’ll focus on higher education, which is largely voluntary and somewhat subject to normal economic principles
  • School to age 16 is largely an organ of the state, and serves social control and indoctrination purposes as much as educational ones.

Before we discuss the technology and its potential effects, let’s take a look at the current system, which to my eyes is already broken.

The first problem with higher education is there’s too much of it.

  • Here in the UK, university participation rates have risen from less than 10% when I went to university (35 years ago now) to more than 50%.
  • There simply aren’t 50% of jobs that require a degree.

Even with the “qualification inflation” of the past 20 years – which means that many jobs which previously did not require a degree now ask for one [1] – 25% of UK graduates don’t find a graduate-level job. [1]My first job after graduation was as a computer programmer, and my intake were the first at the firm to have degrees – I worked in a building of 3,000 people, of whom 6 were graduates

The second problem is that education is too expensive.

  • Costs have soared over the past 30 years, with much of the new spending aimed at administration rather than teaching.

Unlike the US, university fees in the UK are capped (for British students) at £9,000 per year, though that figure will probably be linked to inflation from 2017.

  • For many years fees were paid by the government, and students received a subsistence grant for rent and food.
  • Now everything is funded by a loan, and most students leave college with debts of £20K to £40K, which accrue interest at a favourable rate. [2]0.9% at the moment

These are repaid by a 9% tax surcharge on earnings over £21K (around 80% of average UK earnings).

  • If not repaid after 25 years, they are written off.

We’re already at the stage where the cost-benefit analysis for a degree in certain subjects (mostly arts and social sciences) doesn’t add up.

  • As fees increase and knowledge-based jobs are taken by robots, this situation will only get worse.

When a college degree is no longer an automatic ticket to the middle-class, who will want to pay for one? [3]The third problem, which is outside the scope of today’s article, is that physical universities have ceased to be the challenging institutions of my youth and instead act as “safe spaces” for special snowflake justice warriors driven more by identity politics than a thirst for knowledge


The first potential area of disruption is marking

  • Computers have been marking multiple choice tests for decades, but now they can mark essays too.

A machine marking system is based on machine learning:

  • algorithms are trained using essays that have been graded by human examiners
  • they can then mark new essays
  • samples of the marked essays will also be marked by human examiners to check the success rate

By 2012, the software was as good at marking as human teachers.

Machine grading means that instant feedback can be offered to students while the subject matter it still fresh in their mind.

  • The system can also be used to provide day by day feedback as a student progresses along a course – this would be prohibitively expensive with human professors.

Critics argue that because it is possible to construct nonsense essays that fool the algorithm into awarding high marks, the machine marker doesn’t work.

  • But in fact, constructing a nonsense essay is just as difficult as writing a good one, so this proves nothing. [4]I am reminded of how hard it is to play the piano “badly”, a la Les Dawson or Victor Borge

The point of the marking system is to not award high marks to someone without writing skills, which it seems to achieve.

  • The system does have biases – for example it seems to prefer longer and wordier sentences – but these are inherited from the human markers it was trained from.

The cost savings, speed, objectivity and consistency advantages of machine marking suggest that it will be widely adopted in time, at least on lower level courses.

  • It will be largely graduate teaching assistants rather than full professors who will be displaced by this, and it’s not clear that they have the political power to obstruct change.

In 2013, EdX a provider of free online courses called MOOCs (see below) announced that it would make its essay-grading software freely available to all educational institutions.


The second big potential force for disruption is MOOCs.

Massive Open Online Courses – MOOCs – started in 2011 when two computer scientists at Stanford – famous Google alumni – made their introductory artificial intelligence class available to anyone at no cost over the Internet.

  • 160,000 people from age 10 to 70 signed up for the course

Like most MOOCs, the course consisted of short videos and interactive quizzes to check your learning as you progress.

  • It’s easy to imagine these quiz systems developing into robotic tutors that follow the progress of individual students, varying the pace of the course and offering personalized instruction.

After the course, one of the professors got some venture capital backing and formed Udacity.

  • Two other Stanford professors founded Coursera, partnering with over 100 institutions across the globe.
  • MIT and Harvard got together to form edX.

Here in the UK, elite universities – Cambridge and Oxford in particular – have resisted offering MOOCs.

  • We have FutureLearn, which is basically the Open University, with some courses from other UK, European and Antipodean institutions.

As MOOCs took off, people began to dream of cheap, high-quality education for all, particularly in the developing world.

  • But in practice not many people finish the courses – the average completion rate from one study in 2013 was 4%. [5]Many who enroll never even start the classes
  • And most “student” are already college graduates from a physical institution.

There was also a monetisation issue.

  • Most of the courses were free to “audit”, but you could also buy a “statement of accomplishment” or similar (the wording varies by platform and institution) for a few dollars.
  • Not that these are not “certificates” from the institution (see below).
  • But most people don’t want these, and so the platforms aren’t profitable.

Udacity has largely moved over to short vocational courses, often focused on technical skills like blogging and web development.

  • But these courses, although often relatively low-cost – are written by freelancers rather than professors at elite institutions, and their quality is hard to discern in advance.
  • In any subject area other than the fastest moving, you’d probably be better off reading a good textbook.

Slightly better are the courses backed by tech companies (such as Google) that train people to use their products.

One step up from that, Udacity also has an online Masters in computer science.

  • This is sponsored by AT&T – who will put many of its employees through the course – but still costs $7K.
  • That’s much less than the $35K of a typical degree on-campus, but it’s also a lot to risk on an unproven qualification.
MOOCs – my personal experience

Before I started blogging in 2014, I worked my way through around 50 MOOCs in finance, computers, art and music.

At first I was incredibly enthusiastic and amazed that such high quality information was available for free, but the more courses I completed, the more the cracks showed:

  1. courses are mostly introductory, and it’s frustrating when you want to dig deeper into a subject
  2. the vast size of the classes and the variation across students (in age, geography, language and ability) make it difficult to feel part of a group – it’s basically you doing the course alone on your computer
  3. because courses aren’t recognized as formal qualifications (see below), they aren’t worth paying for [6]I’m not personally in the market for a new career, but for anyone who is, this is a real problem
  4. there’s little variety to the points of view on offer – almost all the courses I took felt like the product of cookie-cutter left-leaning institutions
  5. the courses were generally quite academic and didn’t take advantage of the opportunity to reflect real-world practices

I still try the occasional course now and then, but my personal completion rate has dropped in every year since 2012.


Education is about competition and standing out from the crowd.

  • Colleges don’t sell education, they sell certificates – a piece of paper that says that you went there.
  • Exam certificates are credentials, like passports – they tell someone else (in particular, potential employers) how to treat you.
  • Like paper money, the more of these pieces of paper that exist, the lower the value of each one is.

Elite institutions offer something that can’t be replaced by a MOOC – they have scarcity value and a badge to say that you are part of the in-crowd.

  • So not surprisingly, they don’t want to give certificates to people who haven’t paid, and who might not be who they say they are.

There needs to some solid way of identifying people online so that people don’t make a living taking courses for others.

  • There also need to be automatic checks for plagiarism.

Both approaches could be automated.

  • Plagiarism software already exists.
  • Bio-metric identification (fingerprint and iris scans, facial recognition, typing patterns) could be used to verify identity.

Remote-proctoring (monitoring students by web-cam) has been investigated by some platforms, but this is labour intensive.

Perhaps a better approach is “competency-based education” (CBE).

  • This is a fancy name for what we used to call exams.

Over the past 30 years, education has moved towards coursework and continual assessment as a way of demonstrating achievement.

MOOCs could be combined with open public exams to produce an education system where the method of study was immaterial, and everyone took the same exam on the same day.

  • This method is already widely used for professional qualifications.

This seems like a good fit, but there are commercial considerations.

  • By separating the course from the credential, CBE removes the ability of “elite” institutions to charge a premium price.
  • And cheap or free MOOCs undermine the business model of even non-elite institutions.
The way ahead

Despite being an excellent learning mechanism for a motivated student, and despite their good fit with a world of work that is likely to require continuous re-education rather than a single dollop before your career, the future of MOOCs is currently in doubt.

To mix my metaphors, online learning appears at the moment to be falling between two stools because of a chicken and egg situation.

  • Current “certificates” have no value, and so no-one will pay for the courses.
  • But no existing players have a commercial reason to improved the value of the MOOC certificates.

There is an opportunity for a new company to take on the testing and credential issuing role, leaving students to work out where they will get the information from.

  • This firm would need to build relationships with the employers who are the ultimate customers for the credentials.
  • It might perhaps start in one industry that is particularly open to change – tech, say – and spread out from there.

There might even be a role for government , since the dollar-efficiency of a MOOC based education system is likely to be higher than the current bricks-and-mortar model.

It remains easy to imagine a future where students around the world attend online classes taught by professors at elite universities, in order to receive credentials recognised by elite employers.

  • There would be no need, and no business model, for the non-elite institutions.

But do the elite universities want this, and if not, who will force them down this road?

Until next time.

Don’t forget to share this via , Google+, Pinterest, LinkedIn, Buffer, , Tumblr, Reddit, StumbleUpon and Delicious.

Footnotes   [ + ]

1. My first job after graduation was as a computer programmer, and my intake were the first at the firm to have degrees – I worked in a building of 3,000 people, of whom 6 were graduates
2. 0.9% at the moment
3. The third problem, which is outside the scope of today’s article, is that physical universities have ceased to be the challenging institutions of my youth and instead act as “safe spaces” for special snowflake justice warriors driven more by identity politics than a thirst for knowledge
4. I am reminded of how hard it is to play the piano “badly”, a la Les Dawson or Victor Borge
5. Many who enroll never even start the classes
6. I’m not personally in the market for a new career, but for anyone who is, this is a real problem

Published by Mike Rawson

Mike Rawson has recently re-awoken a long-standing interest in robots and our automated future. He lives in London with a single android - a temperamental vacuum cleaner - but is looking forward to getting more cyborgs soon.

Leave a comment

Your email address will not be published. Required fields are marked *

Disrupting Education

by Mike Rawson time to read: 8 min
Hi there - can I help you with anything?
[Subscribe here]
More in Economics, Editorial, Work

Watson is an IBM supercomputer that uses AI (natural language processing and machine learning) to answer questions and shot to...