With the recent AI boom, there have been many questions about their proper use, what to do with their improper use, and how to handle the situation going forward. Education has been hit particularly hard so far, with teachers and staff facing unprecedented issues.
“Oh, it’s complicated. I don’t think there are any simple black and white answers here, and I think that overall it’s, it’s the Wild West right now,” said Chemistry Teacher Lionel Zhao.
While it may be uncertain now, it is certainly a presence to be addressed, according to a study done by the Pew Research Center this year one-fifth of students who have heard of Chat GPT have used it for schoolwork. Among juniors and seniors, this spiked to a quarter of students. Of these students, 20 percent cite no problem with using AI to write essays prompting concern among educators.
“I am worried they will lose the ability to think critically and write creatively,” said English teacher Jessica Phiri, “I believe more and more students will use AI as an easy way out and will cheat themselves out of an education.”
When asked whether or not they had had a student cheat with AI, the majority either had a definite yes or a belief that it had occurred but no definite proof. Staff seem to be most concerned about AI’s overreliance, frequently citing the fear that students will become dependent on the software.
“AI has made it harder to inspire life-long learners who know how to solve unique and challenging problems and think critically,” said science teacher Richard Wood, “The problem is that access to AI and even access to cell phones and social media is not developmentally appropriate.”
A Harvard study said that despite over 50% of teens having used generative AI only 4% use it daily. Opposing this a quarter of American teachers have found it to do more harm than good. Even when not used to do the work AI can still affect the quality by spreading false information.
“When search engines started pushing AI responses and not filtering results by accuracy and relevancy, it made it harder to help students find reliable sources and critically evaluate information on the internet,” said Wood, “Student motivation to seek out peer-reviewed, validated information is waning.”
Using AI to generate schoolwork and then passing it off as your own is both dishonest and inaccurate, however, the tools used to check for AI work can often be just as inaccurate.
Often, AI work fails to be detected when rearranged by hand or with ‘word flippers,’ and occasionally, even falsely flag work as AI when it isn’t. The latter problem can often be triggered when quoting, paraphrasing, or taking inspiration from pre-existing work.
“I think that if you’re going to come down on a student for using AI, you better be sure beyond a reasonable doubt,” said Zhao.
The best method according to Zhao for finding out real or fake is an oral interview even if it is an inconvenience compared to less reliable AI checkers. Either way, finding AI is only getting harder as students get more familiar with how to use and disguise AI, all the while the software continues to get smarter and smarter.
“At first, Chat GPT was not good at completing complex tasks, but it has adapted and improved its responses faster than I had anticipated,” said Wood, “Some of the first responses to lab questions were D+ or C- answers, but AI can now provide students with B+ or better answers to most questions I would ask on a lab or test.”
Based on this, if AI could consistently perform on this level it would not only pass but even receive honors, with no real work. This is especially alluring to students who might not be able to score this high on their own merit. This said it can be avoided in many classes based on how it’s set up.
“I make it virtually impossible in econ classes. We’ll either do stuff by hand or answer certain questions that, without context, AI can’t do much to help,” said Econ teacher David Brooks, “Econ is more math-like, and again, what we do and how we do it renders AI not as significant as a disruptor.”
This may not be possible for every class but eliminating situations where AI cheating occurs does sidestep the problem for teachers who don’t want to deal with it. AI does not only affect student work, but teachers can also use the software to make their lives easier in many ways.
“I have also tried using AI to write outlines or brainstorm ideas for labs, activities, and lessons. AI helps do some of the heavy lifting on mundane tasks, but it lacks experience and personality, so I still have to go through and modify, improve, proofread, or rewrite anything AI creates,” said Wood.
This faction of using AI on mundane tasks to make lives easier is what makes it so attractive and especially busy students who are juggling school work, internships, jobs, college, extracurriculars, and more can often be enticed by this assistance.
“I think most students are using AI as a pressure relief valve to take things off their plate that are either uninteresting or tedious. If I would have had access to AI in school, that is how I would have used it,” said Wood, “I would have used AI to write the bulk of an English or history essay if I were not interested in it or did not have enough time.”
This, while dishonest, is not a completely unprecedented occurrence, but it is certainly becoming more widespread with AI. It is further fueled by the incredible pressures on students that seem to only grow more and more.
“AI cheating is the democratization of a form of cheating that used to be only accessible to extremely wealthy people. If you had money and you were willing to cheat, you could always pay someone to write your papers for you to write your lab reports for you,” said Zhao, “Now the AI does it for you.”
There is certainly a more moderate use of technology to help students without cheating. The most popular methods are to double-check work, reinstate ideas, and brainstorm.
“It should be done at a level where you gradually introduce it. Then you can rely on it more after you’ve demonstrated that you have the critical thinking capacity yourself to analyze the AI output,” said Zhao. “Isn’t necessarily going to be correct all the time, but it is certainly better than just like you know, thinking it to yourself in your head.”
Students are not the only ones getting in trouble for their often too-liberal use of AI, as the widespread implementation of the new technology has also looked bad for many tech companies due to concerns about the environment.
“Two things in the technology sphere have massively increased the amount of energy that the human race requires to do its civilization. One of them is crypto coins, and then the other one now is AI,” said Zhao,” it is very, very energy-expensive to run AI projects, but good luck telling people that you can’t have the power to do this”
The education system is always adapting to new technology, and often it takes a while to adjust. For example, just this last year the state of Indiana decided to crack down on cell phone usage in the classroom despite them being there for decades beforehand. Many people naturally draw a parallel between these issues.
“When I was a student, teen cell phone use was becoming more prevalent. The argument from teachers at the time was that cell phones had no academic purpose. I have seen the academic value of cell phones in a limited number of applications and with clear expectations,” said Wood, “I similarly see AI. Initially, I was vehemently against AI in the classroom, but I recognize the potential it has to add value.”
The caveat to this is that the use must be proper. In the same way that cell phones are capable of both helping and hurting academic pursuits, AI is as well. In many people’s eyes, it is not currently being used this way.
“I hope that AI can be a positive addition to education and society, but it is not being applied to areas of our lives that could benefit from automation,” said Wood, “I would like to see AI used to improve curricula, create or optimize schedules, accommodate students with disabilities, evaluate student work against a rubric, and increase efficiency by eliminating redundancies in policies and procedures.”
This exact thing was done by Zhao in the last year when he used AI to assist him in standardizing 4-point rubrics throughout school. While he knows how to do the work by himself he found it convenient to use the AI to help. After using the utility firsthand he supports using AI more as a school and a district to standardize policies, as well as create uniform policies on how to handle and address AI.
“I would be in favor of adopting some kind of AI curriculum. There are risks and concerns with it,” said Zhao. “Then again, you know people are using it anyway. Why not make it official?”
The fact of the matter is that AI is here to stay. Despite the many problems, there is also enough utility and investment by both big companies and individuals that there is no going back now.
“Like anything else, from initially having Internet at school to each student having his laptop to now AI, it’ll be a net positive, but it’ll take some learning and adjusting along the way,” said Brooks. “Those who can will thrive, those who can’t will simply continue to complain.”
