Author: Jiajun Zou

PhD Candidate of History at Emory University

It is a memory that all of us have. You submit a paper or assignment to your teacher or professor, who, over the past week or so, has compiled a bunch of papers to read because it is either midterm or final week. From the perspective of educators, their job and passion is to help students as much as they can without allowing work to interfere with their personal life or take away their personal time – maintaining a work-life balance that is hard to achieve yet crucial for sustaining health and work motivation. Thus, there is an unspoken rule in the modern professional world that you are not expected to open your email or answer work-related phone calls after work or during the weekend.

However, this divide between private interest (the need to maintain work-life balance) and public interest (the desire for instant feedback) is at conflict. Learners learn best through instant feedback, even if minimal. Feedback allows the brain to quickly see the blind spots it missed and evaluate the strengths and weaknesses of your work assignment, whether it be homework, an essay, or a presentation. If you bring this up with human educators or professionals, it is likely that they will start complaining about how little they are paid relative to the amount of work they do and how the lack of work-life balance is detrimental to their health and family relationships.

This reminds me of my part-time job at a Chinese restaurant in Queens, New York, where customers liked to order a lot of food at the same time. Friday nights from 4 to 6 pm had more orders than any other time of the week. This rush jammed the kitchen with orders and chefs and busboys started arguing with each other over who was responsible for mistakes, causing further delays and damaging morale. I remember this even after some fifteen years because, at sixteen years old, the restaurant workers thought I would take the blame without consequences. My experience as a customer service agent at a restaurant made me realize just how unhappy it is to receive so much work at once, which increases the probability of making mistakes, downgrades morale, and leads to mediocre service and food quality. If your address is far away, you bet the restaurant delivery person will send your food last even if you call in earlier than other customers. If your surname starts with a Z, then you are likely to be the last one to be graded for a paper. In other words, your surname places you in a position similar to being the last customer of the day, consuming a meal made from the least fresh ingredients, equivalent to an educator’s last bit of mental energy who likely just wants to go out and grab a pint of beer. Your surname that starts with a Z is the final obstacle to overcome before they can leave and grab something to eat or drink.

The reality is, the conflict of interest is an artificial barrier that serves neither the educator nor the students. Educators must be aware of the risks of not providing instant feedback. Students can easily forget what they just wrote, and their brain instantly switches to other important tasks or seeks a drink as a reward for their hard work. Without instant feedback, ideas fade and move to the “storage room” of the brain. When students receive their paper back in a week or a month, they have long forgotten it, and the educators’ feedback no longer feels the same. It’s like a firefighter arriving late to the scene. The fire has already died down, and the owner has moved to a new place. It is a lose-lose situation in which neither side derives any benefits, yet it appears as an unspoken rule in modern academia and professional workplaces. Instant feedback is becoming increasingly difficult to obtain.

In 2023, more than at any other time before. ChatGPT is the first of its kind in which people genuinely feel the benefit of AI in helping to solve their daily tasks. This is different from earlier AI hype in 2016, with Google’s AlphaGo beating the human Go champion Lee Sedol (b. 1983), which caused excitement in the AI community but had no practical benefit for the rest of society. It was a news headline that would make someone nod their head while drinking coffee and say, “interesting.” However, ChatGPT is different. It can help students, workers, businesspeople, lawyers, accountants, and people from all walks of life to jumpstart their productivity—an hour saved from tedious data inputting in Excel is an extra hour to rest or to head to a Korean supermarket to grab some delicious food and drinks. Thus, the divergent attitudes toward AlphaGo and ChatGPT. The former, and later Facebook’s Metaverse, are “optional” and “fun” ways in which AI comes to be known but do not change anything. I don’t play Go and don’t have time for cheap pleasures such as Metaverse. But ChatGPT and LLM technology, in general, can truly transform the human learning experience, especially because of their ability to provide instant feedback. Try it; submit a very long text to ChatGPT and ask it to write an analysis, a summary, or to create a mental map using markdown syntax. No human beings can read as fast as ChatGPT. It’s based on sentence embedding, which allows for quick absorption of your input and instant feedback. You can learn from feedback in a way you never could have imagined.

Should AI chatbots replace human teachers and professors in schools and allow schools to let AI teach students while hiring just the researchers to continue investigating the frontiers of knowledge in science and humanities? Wouldn’t it be a better use of time and mental energy to allow researchers to do what they do best instead of overwhelming them with work at specific times of the year, which results in lower quality and slow feedback? Ah, good luck, students with a surname that starts with Z. You are in a situation that, as the book Freakonomics mentioned, you are waiting for the judge to give you your court decision, but the judge hasn’t eaten breakfast and is depleted of glucose levels in their blood, leading to a lack of empathy and fairness and a higher likelihood of mistakes from your judge.

It seems to me that AI chatbots replacing educators may be a good thing for both sides. First of all, it is the right thing to do. It enables students to receive high-quality feedback, which means they can learn from their mistakes quickly. It is now also possible for AI to have long-term memory of what you say to it, thanks to developments by Pinecone. This means your professor may not remember your name after the semester is over, but your AI chatbot can notice changes in your writing style and ideas compared to ten years ago and tell you precisely which parts are getting stronger and which are not. An AI personal educator that knows about your learning behaviors and patterns will deliver the best educational experience you’ve ever had. There is just one challenge: AI still needs to improve its ability to perform critical thinking and draw connections. However, these challenges are likely to be resolved with time because they are only a matter of computing power, and the marginal cost of computing power will decrease.

In conclusion, human educators are no match for AI and cannot guarantee punctuality, quality, and fairness in their feedback on a student’s work. This will result in a general questioning of the value of school-based education, which consumes most of our lives before we turn 21. Moreover, school education is also a source of great inequality. Your parents’ income determines the neighborhood you live in and, consequently, the school zone you belong to. Whether you grow up well-educated or surrounded by challenges is not up to you, but up to your environment. It is a pitiful situation that artificial barriers are created but go unnoticed or unspoken because humans are afraid of change. If you read Jared Diamond’s book Collapse (2004), societies will not change willingly until there is a crisis, and only those proactive societies can change successfully, whereas the rest are consumed by their own ignorance and deliberate ignoring of problems. For people who value education, they must value instant feedback. Unfortunately, it is not realistic to expect this from human educators, and nor are schools and universities willing to invest their money to improve your experience. It may not be in their interest to do so. People who run administrations and are in a situation of power are typically people older than 35 who are at a stage in which their amygdala is biologically wired to be hypersensitive to changes. Any change from people aged 25 or younger can threaten stability, either as innovation or as a risky disruption of order.

Hopefully, at a minimum, people have already adopted AI chatbots to help with tasks such as grammar correction, punctuation, brainstorming, and receiving instant feedback. Anything that serves in the obtainment and refinement of knowledge is good for education, and becoming resourceful people in a time when resources are not equally distributed and changes are unlikely to come is the best way to control what we can control.

In the future, we may witness a revolution in education as AI chatbots and LLM technologies continue to advance, allowing for personalized learning experiences and instant feedback. This transformation has the potential to break down the barriers of traditional schooling systems, making education more accessible and equitable for all.