Thursday, June 17, 2021

Higher Education is exacerbating, not fixing problems, in teaching - much of it can be automated

The first step in finding a solution is to admit the true nature of the problem. In Higher Education we need to admit that the fundamental problem is that those that teach are not teachers but researchers. Lecturing is easy, teaching is hard and they often don’t have the motivation, skills or time to do what is necessary to teach effectively. Providing tons of Learning Design training and support is not the answer. On the whole this will increase time and costs. It’s time to take a more radical step, and rebalance the system back towards what it used to be.

The current model, the researcher who has to teach, is baked into the system. At its worst you have world class researchers who can’t teach at all. I knew a physicist who was taught by Peter Higgs (Nobel Prize Winner) and described him as “completely incapable of teaching at all”. At Harvard, Erc Mazur described the Nobel winners who teach as “being no better than dinner party commentators when it comes to teaching and education”. Mazur who has spent decades pushing active learning and interactive learning, after shifting his course at Harvard completely online during the pandemic claims...

I have never been able to offer a course of the quality that I’m offering now. I am convinced that there is no way I could do anything close to what I’m doing in person. Online teaching is better than in person.”

But let’s focus back on the core problem, which is the idea that every teacher has to design their own course from scratch. No other area of human endeavour adopts such an individualistic, and therefore costly and inefficient, approach to delivery. Research requires internal intellectual reflection, attention to detail and full analysis, long-form academic writing, with a skew of personality types towards introversion and not great communication skills. Teaching requires strong social skills, emotional intelligence, good communication skills, the ability to simplify content and provide constructive feedback.

Research is the foundation of Higher Education, so what does it say about this problem? Astin’s longitudinal study (1993) on 24,847 students at 309 different institutions, analysed the correlation between ‘faculty orientation towards research’ and ‘student/teaching orientation’. The two were strongly negatively correlated. The student orientation was also negatively related to compensation. “There is a significant institutional price to be paid, in terms of student development, for a very strong faculty emphasis on research” he concluded and recommended stronger leadership to rebalance the incentives and time spent on teaching v research. Note that this does not say that some researchers are not good teachers, nor that some teachers are not good researchers, only that, as Astin claims the research shows that there may be a large statistical mismatch. Note, that this is also not to decry either researchers or teachers, simply to point out that there is a negative correlation. 

Large studies from HEPI and HEA also show student dissatisfaction with the quality of teaching, with around a third reporting disappointment around the teaching, poor interaction, poor communication and interaction with faculty and far too little feedback. One could add to this the evidence that very large numbers of students fail to attend lectures, still the primary pedagogic method in Higher Education globally. Imagine the restaurant business reporting that 40% of their customers failed to turn up for their meals, even after they had paid for them in advance. That would be unthinkable.

How did this happen?

There has been a steady pendulum swing in Higher Education away from teaching towards other activities in terms of funding, status and rewards. Over 50 years ago Jenks and Riesman (1968) showed that salary shifts, promotion and funding suppressed teaching in favour of research. Non-research faculty numbers shrank and, as tracked by Massey & Zemsky (1990) teaching time was literally swapped with research and other professional activities. They described this as ‘output creep’. Boyer (2006) also tracked the year after year swing away from teaching towards research. It was exacerbated in the UK when a large number of Polytechnics were converted into research-based Universities.

The system is now geared towards teaching undergraduates to become post-graduates and faculty, not degree completion. In a profound sense, the research-driven, teaching agenda is punishing teaching and therefore the majority of students. It sometimes exhibits itself as blaming the students, seeing them as the problem. It is rarely discussed but poor teaching puts enormous pressure and stress on students, with high drop-out rates and increasing levels of mental health problems. I have seen plenty of heart-breaking evidence of this.

Breaking the link

Until we break this link, efforts to support faculty with professional learning design will simply plaster over the problem. Recording a bad lecture is a start but not any real solution. However, it may improve even a bad lecture. Yet some Universities still, in this digital age, do not record lectures (my son can testify to this). Yet the advantages are clear; Available 24/7, rewind if your attention drops, rewind if you didn’t understand, rewind if English is your second language, pause if you want to look something up, access to linked online resources on same device, pause to take good notes, fast forward (even 1.25 speed), if known or irrelevant, watch several times for increased retention, watch when in right attentive state for learning, watch if you have been ill, watch for revision as exam approaches, not wasting time travelling to and from lecture, academics can focus on tutoring and feedback, multiple uses in courses, MOOCs etc., data gathered on who, what, when and how long watched, can be subtitled for the deaf, can be translated and subtitled and can be delivered online at any scale, at almost no cost. What’s not to like?

In truth this digitising existing lectures and content into a VLE/LMS was a useful transitionary stage - digitise what you have. We now have to use better tech to improve teaching, not just mimic what we have. The current model, even with the help of Learning Designers, who do a brave and sterling job, is still a process full of resistance and friction. Until we break the ‘'all teachers must be researchers' link, it will continue to be a secondary consideration and learning design support will continue to be a finger in the dam.

Blended learning not Blended teaching

Let’s raise the stakes and bring in Blended Learning. Post-Covid there is a recognition that this should be the norm. Yet Higher Education has a very primitive view of Blended Learning, not as an optimal blend, but as a mixture of offline and online. I call this velcro design, online as an afterthought. This is not actually Blended Learning, it is Blended Teaching. True Blended Learning takes a deep analysis of the learners, types of learning, resources and constraints, then produces an optimal blend. This analysis can and should be automated. If you think that sounds utopian, read on.

Automating Learning Design

We’ve built such a system during the two years of Covid. It takes multiple INPUTS about your learners (distribution, first language, educational attainment, motivations etc), learning (contemporary taxonomy of learning), resources and constraints (available technology, human resources, teaching skills etc). It then uses a state-of-the-art analysis to produce an OUTPUT, which is an optimal blend. Note that this blend may not be a mixture of offline and online. It can be completely online or completely offline. It also produces scores, based on research on success criteria, from engagement to transfer.

More than this we’ve also been automating the production of online content. Using AI techniques that analyse your documents, powerpoints and videos, we can produce online learning in minutes not months. It also links out to external content automatically. This dramatically reduces the time and costs.

Another successful species of automation is adaptive learning, where adaptive courses deliver, not a linear course, but one where students vector through the course at different speeds, ensuring they stay at just the right level of challenge and do not suffer catastrophic failure. These courses are being delivered in the US and China, with proven increases in attainment and correlated falls in drop-out. Going back to my primary argument, why should every teacher invent a course from scratch, when low-cost-per-students online, adaptive courses already exist? The primary reason is, of course, culture. There is no real sharing culture in Higher Education teaching.

Beyond this, contemporary LXP (Learning Experience Platforms) can automate the delivery of learning, pushing and allowing students to pull learning as they progress. Learning is a process not an event. Yet event-based-learning, lectures, dollops of e-learning is what is usually delivered. 


Teaching is a means to an end, not an end in itself. Don’t see the automation of teaching as a threat but as something that frees us to do more research and produce better student results faster, with less mental pressure and at a lower cost. Higher Education should not accept low quality teaching or the description becomes an oxymoron. We now have the technology to hard bake proven, evidence-based pedagogy into the software, to do things human teachers could never do, on scale. We can use AI to personalise learning, deliver the right thing to the right person at the right time and deliver deliberate and spaced practice. My fear is that I see this happening in the US and China but not in the UK or Europe. We have the oldest and finest institutions in the world but that may now be holding us back on teaching. It is time we used smart software to deliver smart learning to make students smarter, faster and at a lower cost. That will increase equity more than any other initiative in Higher Education.


Astin, A. W. (1993). What matters in college?: Four critical years revisited. San Francisco: Jossey-Bass.

HEPI survey (2015)

Jencks, C., & Riesman, D. (1968). The academic revolution. Garden City, N.Y: Doubleday.

Massy, W. F., Zemsky, R., & State Higher Education Executive Officers (U.S.). (1990). The Dynamics of academic productivity: A seminar. Denver, Colo: State Higher Education Executive Officers.

Boyer, E. L., & Carnegie Foundation for the Advancement of Teaching. (1987). College: The undergraduate experience in America. New York: Harper & Row.

No comments: