Thought Leadership
Looking Back to Look Ahead: The Best Way to Think About AI for Education
Head of Learning at Kaplan FinancialArtificial intelligence is fast becoming one of the most important forces of our time. It’s hiding in plain sight within your smartphone, helping Alexa offer up advice, powering your online searches, and targeting advertisements based on your interests. Its latest form, Generative AI, or GAI, has now emerged and been met with equal parts excitement and apprehension – because it brings both opportunities and risks to many sectors.
This is especially true in education, where it has the potential to enhance teacher quality and efficiency, for example, by making grading faster, more consistent, and more accurate. And it can analyse large amounts of data on individual learners so that teaching interventions can be effectively personalised. However, it also raises genuine concerns, such as potential job losses in the sector and perpetuation of bias in the data sets used by AI systems.
Accordingly, we need to better understand the technology and its limitations; more importantly, we must put what is happening into context to ensure we don’t under or over react and safeguard against its misuse. Taking a retrospective look at key developments in learning and education can teach us about what’s happening today.
A very brief history of knowledge
The development of the printing press by Johannes Gutenberg around 1436 is not only one of the most impactful events in learning, but in fact regarded by many as one of the most important inventions in human history: It was the first real knowledge management system, standardising language, improving literacy, and helping disseminate ideas around the world.
However, opinion at the time was not universally positive. First, the printing press disrupted the traditional role of the scribes, previously responsible for copying books by hand. Second, it challenged established religious and political authorities, making possible the dissemination of new ideas and criticisms of the Church. And finally, it enabled the spreading of controversial and false Information, largely because of a lack of controls on what could be said and printed. Sound familiar?
A few hundred years later, in the 1940’s, another technology emerged that changed the way we learn: the modern computer. In 1989, its impact would go into overdrive with the invention of the internet by Tim Berners-Lee. Books were no longer needed to store knowledge, we started communicating online, and the internet effectively opened the doors to the world’s biggest library, all for free (at least for those who didn’t mind giving up their data in exchange!)
But as with the printing press, there were and are concerns as more aspects of our lives are conducted online. Privacy and data security are one concern; the internet also provides a platform for anonymity, which has led to an increase in online harassment, cyberbullying, and hate speech. Misinformation and fake news proliferate, making it difficult to separate truth from lies.
Which brings us to the recent emergence of Generative AI. GAI is a type of artificial intelligence that can create new content: an image, text, or music, based on patterns learned from existing data. The word generative is key: it is not simply finding existing knowledge, as is the case with search, but scanning vast data sets and creating something entirely new. But what makes GAI so special is the fact we can chat with it – and it chats back. To paraphrase Sir Isacc Newton, it stands on the shoulders of giants. The data on which it relies is only there because of the knowledge stored in those early books, amplified by the internet’s ability to capture, and to a certain extent curate, knowledge on a huge scale.
The internet opened the doors to the world's biggest library, all for free.
For knowledge to be of use to any learner it needs first to be curated: selected, organised and maintained. It also needs to be easy to navigate, so that what is most important can be found quickly. FInally, it must be verified, to ensure that the knowledge is accurate, relevant and reliable. GAI does all of these incredibly well, with the exception of accuracy; its current form is prone to “hallucination,” or making stuff up. But over time, this will almost certainly improve.
Impact on education
It’s easy with any new technology to extrapolate its potential to an extreme and then draw conclusions that seem logical but rarely prove to be accurate, due to variables and unknowns. For this reason, in attempting to predict the impact of GAI on education, it’s best to think no further ahead than the next 5, maybe 10 years. As such, the following are simply a few key thoughts to explore possibilities and stimulate discussion as to how GAI can be used for good:
GAI will change our relationship with knowledge.
Each new technology impacts how we think about and engage with knowledge, and GAI is no exception. If we want to find an answer to a question, we Google it. If we want to learn how to do something, we watch a video on YouTube. With GAI, we can engage in conversation, asking questions and receiving answers in real time. This might result in a generation of learners who feel there is no need to learn anything because as Einstein is purported to have said “Never memorise something that you can look up.” But learning is about far more than answers: knowledge is needed to formulate the questions in the first place. Critical thinking isn’t possible without the building blocks of knowledge that underpin the process. What value do humans bring if they simply ask questions without the ability to understand and challenge what they are told? More ideally, learners become experts at formulating challenging and thoughtful questions: seeking answers not just for the sake of a solution but because they lead to a deeper sense of learning.
It has the potential to democratise knowledge.
GAI has the potential to make education far more accessible, because it can teach. It doesn’t simply answer questions; it can be designed to coach, not just providing the solution but encouraging the learner to come up with their own, much as a good teacher would. It can also produce model answers and offer guidance as to how the learner’s answer might be improved. This means that a far greater number of people could have access to the support of a personal tutor, for free.
More do-it-yourself education.
More than 200 years ago Samuel Johnson reportedly said that “lectures were once useful; but now, when all can read, and books are so numerous, lectures are unnecessary.” This ability to self-educate has always existed, made even easier with YouTube and the Internet. Yet technology to date has mostly limited this to bite sized segments of learning driven by the need for a quick solution to a short-term problem. Although the resources are freely available, the motivation and skills to use them are not – nor is there an emotional connection with peers, or a shared learning experience.
GAI, however, can do much of this. You can ask how best to structure a series of topics, putting the easiest ones first. Or maybe you want feedback in a more motivational way, bringing in comparison with others so that you don’t feel alone. Although in practice, getting the technology to do some of the above will be far harder than it sounds, GAI brings the possibility of a do-it-yourself education a whole lot closer.
Teaching and delivery have to evolve.
One important point about GAI is that learning hasn’t changed. It still remains fundamentally about transferring knowledge from short to long term memory – but the teacher now has a very powerful tool to help. One approach would be to show learners how best to use GAI, providing them with the skills to prompt, that is asking the question with sufficient detail and context to get as accurate an answer as possible. It can be a study assistant inside the classroom that the teacher engages with, asking for context to a situation, clarifying a point or offering up a better example.
One final – and perhaps the most critical – challenge: the need to prepare learners to think critically to help discern truth from falsehood amidst the vast sea of information available. We have been trying to perfect this skill for thousands of years and yet it still eludes us. And it’s getting harder as the result of the apparent authenticity of much of the content we see, which GAI can take to a whole new level. Add to this the tendency we all have to seek out and remember only the information that confirms our preexisting beliefs (confirmation bias) and you have a combination that can easily result in division, each side believing they are right with the evidence to prove it.
Challenge: the need to prepare learners to think critically to help discern truth from falsehood amidst the vast sea of information available
Conclusions
As we explore the potential impact of GAI on education, it is essential to consider the lessons of history. By understanding the disruptions and benefits brought about by transformational innovations like the printing press and the internet, we can better prepare for the future. It's crucial to remain vigilant about the use of GAI and to actively promote critical thinking and digital literacy. By drawing from the past, we can harness the potential this technology will bring and create a future where active learning and knowledge are accessible to more people than ever before.
Minority and Low-Income Students are Struggling to Afford College
College tuition and fees have more than doubled over the past 20 years. So it should be no surprise that many college students face a gap between how much their education costs and how much their families can afford.
While the majority of students have unmet need, minority and low-income students struggle with financial gaps at higher rates, according to new data analysis from the Institute for Higher Education Policy. Students with unmet need take out more loans, work more hours, have greater food and housing insecurity and are more likely to drop out of college.
