We live in an era of uncertainty: the gray area between the introduction of a new disruptive technology and its universal adoption. During this time, society is left scrambling for answers on how to manage life with accessible artificial intelligence, for better or worse. And one of AI’s biggest casualties is our education sector.
Could AI permanently change education for the worse? Or could AI be the catalyst of a positive shift in education?
Today, we’ll hear from industry leaders and what they think of the current state of education in the age of AI. Their perspectives range from how AI enhances academic advising to ways of preventing an AI-Driven educational divide. Let’s explore the changes that these experts believe should or should not happen that people may not be aware of.
The Ways AI Is Forever Changing The Education Landscape
To shed light on the potential impact of AI on the modern education system, we've gathered insights from eight experts, including teachers, a senior AI research scientist, a managing attorney, and more.
AI Enhancing Academic Advising
“There has been a lot of discussion about the negative aspects of AI for education, such as concerns about its ability to facilitate plagiarism, or even replace human instructors. However, there are also many potential benefits of AI in education.
In particular, one change that people may not be aware of is that AI can help make academic advising more effective by noticing patterns that might lead a student to fail or drop out. This will allow human advisors to intervene before problems occur. AI may also suggest courses of action, but humans must always apply their own judgment and sense of ethics to ensure optimal results.”
When it comes to AI’s application in education, we’ve become accustomed to hearing about its negative implications. Students getting falsely flagged as AI, overreliance on LLMs, the prevalence of AI hallucination — all of which, we’ve covered extensively in our blog. These are real issues with real implication, and it’s hard to ignore that.
However, there’s proof of what Locklear has mentioned. A recent study from the University of Toronto, a well-known institution for deep learning which spawned AI titans such as Geoffrey Hinton and Ilya Sutskever, tells us that LLMs have unlimited potential in career advising for post-graduate students.
Now, imagine if we could apply that to pre-undergraduate students. In 2019, a Nuffield Foundation study finds that 15% of its respondents are undermatched with their college course, leading to academic underperformance. Using AI as a predictor for the optimal course can close this gap and help students in the long run.
As early adopters of new technology, it’s only natural that we’re hearing about the bad stuff first. But I believe that, in the long run, AI will become an effective driver of education, and academic advising is where we start.
Detecting AI-Assisted Student Work
“As Natural Language Processing AIs continue to improve, teachers will find it increasingly difficult to determine if student papers are written with the aid of AI. It's getting tougher and tougher to tell if a student's work is generated with AI.
To solve this, many teachers have elected to have students write essays on paper during class time. This way, teachers can be sure students do their own work. This approach helps ensure authenticity and originality while keeping students honest about their sources.”
Here at Gold Penguin, we’ve covered almost every AI detector tool on the market. We also regularly test its accuracy against human, AI-generated, and tweaked writing. One of our biggest concerns is how inaccurate it is, and how it can result in innocent students receiving academic sanctions for work they actually did.
The way we see it, it’s a two-way issue. For educators, how can they successfully mitigate the effects of AI in their classrooms? For students, how can they avoid getting falsely accused of using ChatGPT and the like when detectors are inaccurate?
Before the universal adoption of LLMs in education, this issue must really be solved. Callens presents a simple yet elegant way of circumventing the problem. However, as more industries shift to a hybrid setup, his method may become harder to enforce in the future.
Balancing AI and Emotional Intelligence
“On one hand, AI brings incredible efficiency and personalized learning opportunities, but on the other, it risks creating a one-size-fits-all approach, potentially overlooking the nuanced needs of individual students.
In my experience, emotional intelligence is essential in learning, something that AI might struggle to replicate or address. A key change I'd suggest is integrating AI tools with a strong emphasis on developing emotional and social skills. This is about ensuring and maintaining a balance where technology complements the human aspects of education such as empathy, creativity, and interpersonal connections.”
— Bayu Prihandito, a certified psychology expert, life coach, and the founder of Life Architekture.
For years, the Turing Test was the industry standard in measuring behavioral intelligence in AI. Some scientists argue that ChatGPT has already surpassed it. Some also argue that a chatbot from the 1960s called Eliza beat ChatGPT in a Turing Test.
The fact of the matter is, there’s no reliable way of measuring intelligence in AI, much less emotional intelligence. If AI is the future of education, then there must be a way to integrate or mimic emotional capacity in their systems. We can’t have the next generation of students be taught without empathy.
Disruption Requiring New Skills Focus
“Generative AI and other AI-powered writing tools will bring both benefits and challenges to the classroom, and I have mixed feelings about this. I mean, perspectives on the impacts of platforms like ChatGPT in education are all interesting and valid. But all I know is that AI will definitely disrupt the modern education system.
For example, AI can quickly write an essay about a book for a student. But then, is it really proficiency on the student's part? Future education might need to shift its focus more towards complex analysis and interpretation of perspectives, demanding deeper critical thinking. Educators will need to concentrate on teaching and evaluating skills that AI struggles to mimic, such as critical thinking, creativity, communication, and problem-solving.
With the advent of advanced tools like GPT-4, educators will face significant challenges in adapting. It will be fascinating to see how this technology impacts the methods schools use to evaluate student performance.”
AGI is a significant milestone in artificial intelligence, but students were quick to catch on its biggest flaw: unregulated access. Because of this, students can now create essays or code that would normally take days in just a single prompt.
This presents a new challenge wherein we can’t accurately evaluate a student’s proficiency. Search engines are one thing, but to have a chatbot give you an accurate and articulate answer is a completely different ballpark. As Ochoa suggests, there must be an emphasis on assignments that require more critical thinking in the future.
Viewing Potential and Risks
“AI has significant potential to transform education, but we must thoughtfully shape its implementation. AI tutoring systems that adapt to students' needs could provide more personalized and equitable learning. However, if schools become overly reliant on AI, it may undermine the roles of human teachers.
We must balance technology with the vital socio-emotional skills that teachers cultivate. AI should augment quality instruction, not replace teachers. Another risk is data privacy—collecting student analytics risks mishandling their personal information. Strict safeguards are needed to prevent misuse.
We must also correct AI bias that could propagate unfair treatment. With vigilant governance, AI presents huge upsides, like enabling customized education at scale. But it requires development aligned to human values, not just capabilities. We have an obligation to guide AI's emergence in constructive ways. If done responsibly, AI may profoundly improve educational opportunities.”
Every new technology comes with its own set of risks. It just so happens that, with AI, that comes in one of the most important aspects of society: education.
Apart from what Shaurya has mentioned, we should also consider the following risks of AI in education:
- Demographic Bias. Since LLMs are trained on a large and standardized dataset, biases against a certain demographic could emerge if they are significant enough in the training set. For instance, a Stanford study finds that AI detectors are more likely to reach a false positive conclusion for essays created by non-native English speakers.
- Data Ownership. Who owns the data contained in the training set? How are their data protected? This is significant enough that some authors consider suing OpenAI for copyright infringement.
- Loss of Expertise. Dependence on AI could result in educators and students losing touch with their knowledge and expertise over time.
Endangering via Overreliance
“The impact of AI on the modern education system is a double-edged sword. While it promises personalized learning and efficiency, it risks widening the digital divide and diminishing the irreplaceable human elements in education.
The critical change that needs careful navigation is maintaining a balance between AI and human interaction. Overemphasis on AI's technological advantages could overshadow the essential roles of empathy, ethical guidance, and social engagement that educators provide.
It's crucial to recognize that AI should augment, not replace, the human touch in education, ensuring that the system remains inclusive and holistic, rather than purely data-driven and impersonal.”
According to an article by Forbes earlier this year, around 50% of college students use ChatGPT to complete their assignments and write essays. Even if we assume that only a small fraction of that keeps coming back to ChatGPT, it’s still a significant enough number that educators are rightly worried about students’ overreliance on LLMs.
We all know that abstinence doesn’t work. Instead, there must be an emphasis on teaching students how to properly use AI as a tool for their workflow. We cannot prevent artificial intelligence, so responsible adoption is the only way forward.
Preserving Human Interaction in AI Integration
“AI could significantly impact the modern education system by potentially replacing traditional teaching methods with personalized learning experiences. However, one change that should be approached with caution is the complete replacement of human interaction.
The value of teacher-student relationships and peer-learning experiences is irreplaceable. AI should be used to enhance and support these relationships, not replace them. The focus should be on a balanced integration where AI aids learning without undermining the essential human elements of education.”
— Einav Biri, CEO of FARUZO
There are a lot of studies that support the impact of human interaction in education, even before the rise of AI. For instance, a 2013 study from the Western Michigan University highlights that having active engagement fosters a “positive classroom environment and [establishes] a community of learners who support each other.”
The crux of the matter is that, even as society and technology evolves, we must not lose the human factor in education. As Biri suggests, this is an essential element in teaching, perhaps the most important.
Preventing an AI-Driven Educational Divide
“AI-powered educational systems might end up being more expensive than traditional methods, creating a divide where only the wealthy can afford cutting-edge technology and access the best teachers. This sets the stage for a two-tiered education system, leaving those with fewer resources behind.
This potential scenario underscores the importance of addressing not just the benefits but also the societal implications of AI in education. We must advocate for policies and initiatives that ensure equitable access to AI-powered educational tools, preventing the formation of a system where the quality of education is determined by financial means.”
Across the globe, world leaders have been discussing proper AI governance laws which could help mitigate its impact not just on education, but on society in general. Even OpenAI, the company behind ChatGPT, is taking precautions to prevent overreliance on AI.
Beam suggests that AI could create a divide wherein students with less resources are left behind. However, this may not be a significant issue moving forward if OpenAI continues to be the most dominant force in the field and keeps their ultimate goal of creating a universally beneficial AGI. And, with Altman back at the helm after his rollercoaster firing in November, there’s no reason to think they’re backing out of this promise.
To Put It Simply
Let’s go back to the question asked in the title: How will AI destroy the modern day education system?
The answer might not be as insidious as you think.
AI will destroy modern day education to make way for the education of tomorrow.
Will AI become a net negative for education? Time will tell, but I don’t think so. 40 years ago, students didn’t use the internet. 30 years ago, there were no search engines. Education and technology are so intrinsically linked that the former adjusts to the whims of the latter. Whenever technology evolves, so does education.
Modern day education will be obsolete in five years. The shape of what’s to come is yet to be determined, but I’m excited for where AI takes us.