On one random day, I received a WhatsApp message from a tech company requesting that I apply for some remote job.
At first glance, I couldn’t make head or tail of the offer.
They seemed to ask me if I was willing to use my knowledge in chemistry to help them build a large language model (LLM).
Their pay was not top notch but it was reasonable.
So I signed up for it and within no time, I was on my laptop designing ‘prompts’ to train these large language or AI models.
So what was I doing on a day to day basis?
Well, I was required to come up with extremely complex scenario-based questions that could ‘beat’ these models being ‘trained’.
At first, most of my prompts hit a brick wall.
Questions I had initially thought as tough were easy pizzy.
Infact I was discouraged from using the internet for these reasons.
So I defaulted to complex text book questions with some added flavor and it started to work.
The model started missing the mark and getting it wrong.
Getting it wrong was the step in the right direction.
Why?
Because the goal was to ‘train’ it to answer those wrong questions in the long run.
At first, I didn’t understand the gist of the work.
But later on I realized that these guys were actually preparing an educational AI model for deployment.
And this wasn’t just any tech company.
It was one of the big 5 globally.
The company I was working for had just been sub-contracted.
It just occurred to me that we are at the cusp of a major paradigm shift in the higher education sector.
So as I sat down with this realisation, several questions raced in my mind…
1. If students could have access to such robust educational AI models, then what is the future of the university lecture hall?
2. What is the future of examinations if a student gets all the answers on a click of a button?
3. What’s the future of the degree certificate?
I posed these questions (or related ones) to ChatGPT and its response prompted this article.
EVOLUTION OF HIGHER EDUCATION
First off, predicting the future is as uncertain as creating an unknown being.
We are limited by our own understanding, biases and finite intelligence.
With the advent of Artificial General Intelligence (AGI) and super intelligent models, our grasp of the future could at best be a race of “playing catch up”.
That being said, what options do we have as players in higher education?
As we speak, many universities are doing things business-as-usual.
Students are reporting, lectures are going on, exams are being prepared and administered – it’s the normal mixed bag of goodies (or baddies) students sign up for.
But then, there’s the AI disruption that is coming along – or should I say, leaping ahead of us.
Assignments and projects are small time ‘inconveniences’ that can be dispensed with within a minute or two.
Infact, AI scholar models could list for the students credible references that go with the whole ‘bag of goods’.
So if uni exists for grades, graduation and degrees alone, then we’re in for a rough and bumpy ride.
But what recourse do we have as educators?
Truth is there’s no one-size-fits-all answer.
But are there indicators that could help point us towards what’s on the horizon?
Probably.
INTEGRATING AI IN TEACHING
What if I were to do what I do but do it faster and more efficiently?
That’s what integrating AI in teaching does.
AI not only helps in us in creating good educational content but also setting up high quality assessments for the students.
Oh yes. Students also have access to these tools.
But as subject experts, we know what to look for – and what to ask – which puts us a step ahead.
But the thing with progress in this age is that it’s so fast – and it’s bound to get faster as we move on.
Sooner or later AI models will become good if not better student tutors.
Two students could have custom-made lectures in an identical subject area but delivered so specifically to suit their different needs.
Welcome to the age of agentic AI – models that act more or less as your personal assistant.
For universities to survive this period, then they’ll need to rethink examinations and grading.
A new approach would be necessary otherwise lecture halls will be emptied by bots that deliver smart learning to students at the comfort of their beds.
With technology moving ahead fast and furiously, our only asset will be a narrow time window before AI catches up and leaves us panting.
Now this is where problem-based learning comes to play.
By this period (probably 10-20 years from today), AI models could be so robust that not even the best professors will keep up.
However, its possible that robot deployment would still be in its nascent stages.
And that’s a loophole we’ll need to take advantage of.
PROBLEM BASED APPROACH
Cutting edge universities could convert their lectures halls, labs, workshops and libraries into centres of excellence in AI assisted research into common problems facing us.
Climate change, water scarcity, cancer treatment, poverty alleviation, renewable energy, food production and storage – name it.
A typical student will dabble up as a research assistant coming to a university not to gather knowledge but to use its infrastructure to solve these wicked problems.
Since AI provides distilled solutions, human hands could be used to administer these solutions to the ground in the form of projects.
So a student attending university could write a practical proposal to use the academic infrastructure to dig deeper into problem solving.
They could then qualify for funding – or come with their own just like paying school fees – allowing them to research into, iterate and launch solutions to the world making it better a step at a time.
But this time, most instructions, interactions and even mentorship will be done by virtual AI models tailored for those universities according to their areas of interest.
The human educator will evolve into a guide, a mentor in knowledge management, an ethicist, possibly a philosopher and an empathetic voice.
Believe you me, human lecturers will be useful.
However, not all of them.
Only the ones who have superior soft skills and emotional intelligence that AI may – at that moment – be deficient in.
Fast forward 10 – 20 years from this time and agentic robots could make it to the scene.
SO WHAT NEXT?
The age of humanoid robots will mark a significant shift in the delivery of higher education.
This is no sci-fi talk. It’s real business with entrepreneurs like Elon Musk leading the way today.
Robots could be deployed to work on the most manual of tasks – especially repetitive tasks.
Trade skills will have been resilient this far and thriving.
Your plumber, mechanic, electrician, mason, hair dresser would still be making it big in business but not for long.
Self driving cars could also become the order of the day.
AI automation could replace the human being in machines, earth movers, tractors, longhaul vehicles, planes, ships, cranes, name it.
Even the problem-based research output from uni could require less human input – if any.
Picture this:
Imagine an AI bot designing a highly complex model to tackle climate emergencies in an urban slum in an Africa city.
After crafting solutions, it could need real hands and boots on the ground to do some manual engineering jobs.
The same tractors and earth movers we use everyday could now be deployed but act autonomously.
Waterways will be diverted, houses reconstructed, soil could be moved, barriers could be erected, systems could be installed with laser sharp precision – all without the involvement of human hands.
So if we combine AGI, super intelligence and advanced robotics, the human factor could be eliminated to a large extent.
Infact, AI models will even be able to map out and launch businesses without the involvement of humans.
They will be able to run manufacturing lines, distribution channels, marketing campaigns and even boost the bottom line without us.
Now this will be the tipping point not only in education but in human life.
What’s in it for us?
A typical university during this age – probably 50+ years from today – may run using a lean human staff.
Most – if not all – systems will be run and managed by an AI architecture so advanced that even the most intelligent professors cannot match up.
But possibly, an encouraging thing is that human-based skills – empathy, genuine laughter, facial expressions, genuine conversations – will be in short supply.
It is possible that a two-tier system might emerge during this period.
A top-tier composed of people who have integrated AI into their daily lives – and by this I don’t mean having a smart phone but having brain implants or peripheral neural networks that link brains to AI models.
These could be the super-human cyborgs of the age delivering efficiency in everything they do like nothing else.
But on the other hand, a lower tier could exist made up of humans who have chosen the traditional way of doing things.
They’ll probably stick to – at most – mobile phones, and choose to work work with other humans more than bots.
Some universities could have students who prefer that human face – students who prefer a real human talk, gesturing, cracking a joke, sighing, get tired, get emotional and so on.
Probably life might be better that way.
But for the most part, machines will run the show.
The future university will be a microcosm of what’s happening in the real world.
Probably many human-based occupations could flourish during this age – psychologists, counsellors, mentors, coaches, motivational speakers, encouragers etc.
But all of them could be based on the lower tier because that’s the arena where humans will be playing in.
And this is assuming that no malevolent actors take advantage of technology to harm humans. God forbid.
Bottom line is, invest in human skills because a time is coming when they’ll be golden.
