The curriculum for a well-educated Victorian was much simpler than a student of today: There would have been Latin and mathematics to learn, of course, along with geography and history. But when it comes to a subject like science, the differences between what was possible to learn in the early 19th century—when science was dominated by religion—and today is staggering.
Evolutionary biology was yet to have its impact, and Newtonian physics was the only physics that would have been studied. Today, we take genomics and the quantum world for granted, but these were a far cry from the challenges of Victorian life.
You don’t need me to tell you that a new wave of technological expansion is upon us. Experts agree that this period of rapid change will require a different set of knowledge, skills, competencies, abilities, and characteristics. But those same experts can’t seem to agree on what they’ll be. There is little, if any, consensus about exactly what our students, employees, entrepreneurs, and leaders will need to know and be able to do in the future. The result is an inevitable uncertainty that makes planning for education a particular challenge.
Technology contributed to this information overload. But it’ll also help us understand it.
Each thing we learn about the world that we didn’t know yesterday gets added to an ever-growing reading list of “things we should know.” Take a four-year undergraduate computer science degree, for example. In 1980, it could cover most of what needed to be known about computer science and software engineering. But today, while the theoretical core of computer science is largely unchanged, computer architectures have developed enormously and have therefore caused the potential knowledge areas to balloon: parallel and distributed systems, networks, security, cryptography, the list goes on. Then there are all the different applications of computing, from graphics to scientific and medical computing, or human-computer interaction and AI.
Thanks in part to some of those early computer-science students, we have now built AIs that can outstrip our human capability to accurately store and recall information. Beyond AlphaGo beating Lee Sedol at Go, IBM’s Watson winning at Jeopardy, and AlphaZero teaching itself to win at chess, Moorfields Hospital and Deepmind have trained an AI to identify pathology automatically, and we have built AI tutoring systems that can perform at human level.
These AIs continue to learn day and night as they process new data. They continually absorb new information that, when combined with their existing knowledge and the right questions from their human operators, increases their capacity to answer questions and solve problems.
They’ll also force us to redefine our outmoded definitions of intelligence. In this future, intelligence is a concept that is far richer than it was previously. Instead of being seen as all about academic knowledge or IQ tests, intelligence will be seen as a networked, collective synthesis that involves both human and artificial intellects working complementarily together.
We can use these technologies as educational tools.
What will this look like in college? Academic institutions will use AI as an infrastructure to support the evolution of this new intelligence. They will develop complex, sophisticated, artificially intelligent systems that can help us parse that new world.
An enormous range of different technologies are already appearing on the education market, such as augmented-reality headsets, virtual-reality simulation, voice-activated personal assistants, and personalized tutoring systems. These technologies can enable us to interpret student data appropriately and insightfully, ultimately allowing us to support teachers.
- Research projects like MaTHiSiS are exploring how educational platforms can provide every learner, whatever their setting or device, a bespoke, individualized learning experience that is adapted to their personal requirements.
- AI systems like CenturyTech and Alelo use big data and AI to provide engaging personalized learning for students and detailed feedback for educators.
- Technology platforms such as those developed by Founder4Schools use machine learning to help teachers match students to work experience and career sessions with the companies that are developing the fastest in their local region.
Then there are AIs that can augment human teachers in very useful ways. I have previously written about Colin, the AI teaching assistant used by college teacher Jude. Colin is not yet real, although all the technologies needed to create Colin exist and merely need to be assembled, implemented, and trained.
Colin isn’t a robot: It’s an abstract teaching assistant that, like Siri or Alexa, can be accessed via a voice interface. The teacher becomes a metaphorical judo master, harnessing the data and analytical power of her AI assistant to tailor a new kind of education to each of her students.
“The students have already completed some activities in advance, and Colin has integrated their performance on these activities into each student’s personal profile. Jude can therefore see where students are struggling to understand the concepts they need to grasp and where there is greater comprehension. The term performance is now much richer than a decade ago, when Jude had to concern herself with course grades, student satisfaction, and which students might be about to drop out.
Nowadays, the multiple data sources harvested from across campus, from social media, and from students’ online interactions are integrated with student-submitted information, tutor-submitted information, and the latest workforce skills data. At any moment, Jude can find out about each and every student in terms of their knowledge, understanding, skills, well-being, metacognition, critical ability, and self-awareness. She can find out about her whole cohort, sub-groups, compare them to other similar cohorts, and map them to the latest data about what the workforce requires. It’s a far cry from the exams, lab reports, and essay marking of the old days, that could be gamed and plagiarized.”
Of course, these technologies are expensive to develop: A fully functional Colin would be a development project that would cost multiples of hundreds of millions of dollars. Once the initial developments costs have been found, the ongoing production will be much cheaper, but that doesn’t stop the barrier to entry being exclusively high.
The choice of which new educational technology to use is mind blowing, but the price-tags are often budget-blowing, too. This leaves many of the most disadvantaged students behind.
For example, in the developed world, 94% of young people aged 15 to 24 use the internet compared with 67% in developing countries, and only 30% in the least-developed countries. Most worryingly, almost nine out of ten unconnected youths live in Africa or Asia and the Pacific. The evidence about the efficacy and appropriateness of these new technologies is also currently largely missing. This will require concerted effort and investment to correct.
We have built technology that could, for the first time in the world, enable us to provide an education to every learner. The implications of this for balancing inequalities is amazing. However, if we want this to be the future, then we need to ensure that we use our technologies to help us understand ourselves as learners, so that we can become smarter alongside our peers, both artificial and human.