As we enter a new era dominated by artificial intelligence tools, we cannot afford to turn a blind eye, nor to resist change. Failing to thoughtfully incorporate ChatGPT into what is bound to become the new generation of education systems would likely have both pervasive and unintended consequences.
On April 1st, the Italian data protection watchdog, the Garante della Protezione dei Dati Personali (Garante from now onwards), came under the spotlight following the decision to temporarily ban ChatGPT on devices connecting from Italy, being the first Western country to do so. That was no April’s fool prank as many may had hoped for.
Interviewed by Corriere Della Sera, Pasquale Stanzione, head of Garante, remained vague. He pointed the finger at “the plurality, the convergence, and the gravity”1 of the implications ChatGPT could have on minors both in terms of the reliability of the chatbot’s responses, and the collections of minors’ data to train the algorithm2. Fearing other countries could also ban OpenAI, ChatGPT developer company, has accepted to reassure the Garante by improving the platform. By the time you are reading you should be again able to use ChatGPT.
In the meantime, the European Commission is actively working on a comprehensive legal framework to regulate AI systems. Although some applications of AI will be fully banned as they pose “unacceptable” risks3, it is reasonable to believe that ChatGPT will be allowed to operate. Soon it will be up to us, starting with education, to determine how best to use this technology.
First reaction: “Shock“
When ChatGPT was launched in November last year the first reaction among many educators was in unison: pure panic. One could argue that roughly no one, except for some experts, had foreseen how quickly the AI-powered platform would spread. In the first two months alone, ChatGPT is estimated to have reached 100 million users, making it the fastest-growing consumer app in history4.
Students nowadays can request ChatGPT to write cogent essays, complete assignments or find test answers without having to extricate themselves through the web. Occasionally the algorithm may blatantly provide fictitious answers, yet it performs impressively well in a variety of tasks, and it is getting better.5 ChatGPT-4, the latest version of ChatGPT, has a 40% higher chance to produce factual responses and when tested on the bar exam it scored higher than 90% of aspiring lawyers6.
Thus, that AI stands to reshape the normality we are used to should not be up for debate and, as worrying as it may be, we should take ChatGPT for what it is: the tip of the iceberg. Last month tech leaders, including Elon Musk, called for a six-month halt to the creation of advanced AI technologies,7 fearing that further advancements in the sector could spiral out of control. Unlike past technological revolutions which ended up empowering the human species, this time AI may take power away from us as we risk becoming over-reliant on it.
The instinctive rection in most schools was to promptly ban the AI-powered chatbot as it was launched without warning amid the school year. ChatGPT has suddenly made cheating easier than ever, and educators stressed like never before. So no wonder it was met with hostility by teachers.
The Luddite temptation
Fortunately, this is not the first time societies are compelled to adapt to sweeping technological advancements. But despite history showing how market forces in the end prevail, the temptation to resist the spread of disruptive technologies remains inherently human. As far back as 1818, a radical movement of English textile workers, the Luddites, insurged against the use of machines in their industry. Not only did the new technology pose an immediate threat to their jobs, but it also put into question the past normality they were familiar with8.
That sentiment of disorientation and uncertainty will once again become more evident. Take the case of a high school teacher in the US who asked ChatGPT to evaluate his students’ papers. He admitted that ChatGPT was able to give more detailed and useful feedback than he would have within just a few seconds9. Worryingly, a growing amount of evidence shows that the frustrating implications of AI are profound and without precedent.
AI is moving its first steps, yet it is already proving that it can outperform humans in professions we believe could never be delegated to machines. This needs not to imply that we will eventually see white collar jobs, educators for example, being replaced by artificial intelligence. Still, professions must adapt to a new era powered by AI to be able to control its implications and exploit its far-reaching potential. Giving in to the “Luddite temptation”, for example by blocking access to ChatGPT, could be disastrous.
For starters, banning altogether internet access to ChatGPT would quickly prove to be easier said than done. Because any ban can be readily circumvented by accessible VPN services, students who secretly exploit ChatGPT’s potential would end up being evaluated with the same criteria as those that do not. And although plagiarism detection tools may become more accurate at spotting AI-generated text, as ChatGPT learns to sound more human and sharpens its eloquence it may also get harder to detect.
We cannot therefore afford to blindly resist the changes AI will have on education systems and we cannot let business run as usual. In the years to come, the challenge must be to ensure that new generations are best equipped to minimize the unintended implications of artificial intelligence while taking full advantage of the opportunities it presents. To achieve this, it is crucial we understand where the impact of ChatGPT is already tangible, and which adjustments need to be made. First, however, we must update our traditional notions of what it means to be a student or a teacher.
What the “New Education” will look like is up for us to decide
Forget going to the library to borrow a book to write your thesis or browsing around countless hours to find good sources for research. Instead, ask ChatGPT what it can do to help you become a more efficient, original, and employable student. The chatbot will nimbly fulfill almost any task, from summarizing a topic covered in a lengthy book to preparing you for a test with personalized mock exams, or helping you learn a new language.
Given these premises, teachers must reconsider how to measure merit and evaluate classrooms’ performance. Take-home assignments, for example, must be reimagined if they are to retain their meaningfulness. Educators are already experimenting with innovative ways to do so. Recently some students at Mississippi University were asked to use ChatGPT to generate a written argument on a particular issue and then were challenged to revise it fixing ChatGPT’s shortcomings10. Similar approaches build on the idea that ChatGPT should be treated as a sparring partner with which students can sharpen their judgement and opinions.
Universities should channel their efforts to expand students’ curriculums to include a mandatory course that fosters AI literacy as well as critical thinking abilities. Bocconi already provides the basics of Excel and Python to first-year students and could therefore pave the way for further technology-oriented education. From now onward it will be crucial to have a solid understanding of the functioning behind AI algorithms and what broader consequences may arise from their use as they become part of our lives.
As ChatGPT makes its way into education systems, we will get a deeper understanding of which adaptations are needed to ensure we take full advantage of the chatbot. Meanwhile, AI-powered applications will continue to proliferate, irrespective of temporary bans or the opinions of tech leaders such as Elon Musk. After the impact the pandemic had on education, teaching at school and university will undergo yet another transformation. It is up to us to determine the outlook of the new generation of education systems. Unfortunately, ChatGPT has no suggestions for what it might look like, just yet.