Leveling Up Learning (and Grades!) using AI
Don't be an AI Addict or a Stay-Away: be the AI Master!
Happy 2025 New Years! With the turn of the years and the completion of first semester midterms, every student is looking for ways to speed up their schoolwork and put it into practice a very common proverb: “work smarter, not harder”. Unfortunately, it’s really hard to make this a reality. After all, a chemistry lab is a chemistry lab; a ten page Macbeth essay is a ten page Macbeth essay, right? So how do we speed up our work while not sacrificing accuracy?
Let me tell you how it worked for me. Before we dive into that, it’s important to understand how the emergence of AI amongst the backdrop of the pandemic and school shaped how students behave today.
2022, the Year of two Cs: Covid and ChatGPT
In November of 2022, ChatGPT released to the public, taking the world by storm. We had an AI assistant at our finger tips, seemingly with infinite knowledge and capabilities! It was not long before students either hacked their parents GPT account or made their own to start “helping” them in school.
As a consequence of the massive online-to-physical shift that had taken place a year ago, most student materials were on online platforms, the most common one in the US being Schoology. English classes particularly had drastically reduced the number of handwritten, on-paper essays in class and increased the amount of typed, long-term essays that were to be submitted as a google doc to the assignment.
Starting, much less writing, an academic essay is hard on paper: it’s even harder online. Did you know a handwritten draft helps you brainstorm ideas easier? Thing is, no one actually cared. The goal was fast, fast, fast; 100%, 100%, 100%. Now guess what tool is “fast, fast, fast” and SUPER accurate?
That’s right: AI
In less than a month, students began to use AI in massive numbers, almost proliferating prompts everywhere, at anytime. Find a quote that tells you why Lady Macbeth is evil? Instead of flipping though your 100-page Shakespeare text, just ask GPT! How do I balance a redox reaction? Too lazy to go to the textbook, just quickly ask GPT! Forgot how to perform a systematic search of a binary tree in C++? No more hours on StackOverflow, just ask GPT to write it for you!
AI ≠ ChatGPT
It’s interesting to note that although I reference ChatGPT a lot here, it isn’t my favorite model: indeed, you might have not heard about anything other than ChatGPT, but many exclusive, fast, and all-powerful models exist. I use them in tandem with each other, getting two responses from different models which then I can use a little from each to craft my product. Some of my favorite ones include Elon Musks’s xAI, Perplexity, Llama, Gemini, and Claude (all of these are LLMs— “Large Language Models”).
Soon, the school meta wasn’t memorizing slides or finding the secret textbook where the test problems came from. With all semblance of “hard work” out the window, we students asked ourselves one question:
“Who here can ChatGPT the fastest?”
However, clearly we forgot to ChatGPT Newton’s third law: for every action exists an equal and opposite reaction. And this opposite reaction to the surge of ChatGPT was the increasing occurrences of plagiarism. Indeed, students could now just copypaste their essay prompt and copypaste that answer back and submit. For better or worse, two things happened: teachers started monitoring AI work and stated catching cheaters, and second, students’ performance on tests deteriorated rapidly. Because everyone was GPTing their Ancient Civilizations Project, they were failing test questions like “Name the ancient civilizations”.
Tit-for-tat
So, the catch of using AI is that students tend to become overreliant on it. For us, the short-term goal of “overcoming” a mountain of homework outweighs any risk of not learning the material fully enough: we either tell ourselves that we can prepare for the test later and give excuses about what will actually BE on the test (best case), or we give ourselves a false sense of security when receiving back an A+ grade on a ChatGPT-ed assignment that we know the material (worse case). But for now, students were cruising along with their beautiful new engine—both literally and figuratively.
The Addicts and the Stay-Aways
As this AI-centered meta began to develop and become normalized in school culture, especially high school, the initial excitement began to dissipate. Suddenly, high school assignments weren’t straight forward: instead of balancing 20 chemical equations on a worksheet, you had to perform a Beer’s law lab with classroom materials? Instead of solving pages 110-117 in the Precalculus textbook, you had to create a graphical art on Mathematica?
The School Competitive Scene
Overall, two main categories of people developed in perspective to AI: the addicts and the stay-aways. The former were the ones who had become overreliant and attached to ChatGPT. They used it like a fruit, squeezing all the juice they could get out of it and completely abandoning anything given in school. These students also copypasted the whole answer from ChatGPT and didn’t edit it at all, almost ritualistically, wanting to keep the “accuracy” of the prompt intact. They lost confidence in themselves and turned to AI as a walking stick rather than a street-light. Although these students might get a good grade on a few homework worksheets, they often had concerning knowledge gaps come test time. On the other hand, we have the stay-aways. These were students who had been affected very little by AI and had developed an intrinsic repulsion to AI. Although it sounds noble, these people often have an account for small questions but otherwise toil and trouble for their assignments. They are still less proliferating then the addicts, but ironically, still perform suboptimally on tests, complaining about how bad the study materials were. These students often don’t seek to close their knowledge gaps as they wear a cloak of reassurance, comforting themselves that everything the teacher taught would be on the test and nothing else.
Like the world in general, however, these groups were not black-and-white. Like a pH scale, students range anywhere in between the addicts (red and acidic) and the stay-aways (purple and alkaline). I can clearly place many of my friends on this scale as we continued high school.
What about me?
As for myself, I hate to admit it, I was more on the addict side, like a diluted lemon juice. My ChatGPT account was full with all sorts of stuff, with channels for everything under the sun. Luckily, I had the foresight to actually study for the test because I knew I didn’t know everything, albeit the night before. Ironically, I used ChatGPT practice tests. After a few months of using this tool, my mind began to picture how a perfect AI would look for my needs. I would like to think it was my school brain coming together to look and help me in the long term, but it was probably my greedy mind wanting an easier way to use AI while not falling into the pitfalls of overuse.
Revelations and Recommendations to fellow high schoolers
I made a few subconscious observations about how AI behaves, and it shaped my revelations and learning of how to use it effectively. Here are three quick facts and tips about AI:
Not gonna lie, it’s like a tricky genie, always trying to find a way to give you something weird and unintended but still technically your wish (like if you wish to be a millionaire, the genie might give you a million Iranian Rial, equaling 23.75 dollars). Every additional word you add to that prompt increases the chance of it following the specific directions and giving you a valuable, usable answer.
Read its answer! Too many times, people just blindly copy-paste its answer and then later, during presentation time, realize it put something blatantly incorrect/inappropriate for the assignment. And don’t just skim: read each and every word. Check for inaccuracies, and DON’T HESITATE to ask it something else to edit its answer/revise it on your own. No need to worship its first answer. Remember, this isn’t a 90s phone call in which every minute online costs something.
Don’t fuddle the channel. Many of my water-neutral friends use AI for a quick question here-and-there, but they only have a single channel in which they dump everything. For example, yesterday they were asking about the electron configuration in Boron Trifluoride, and today they’ll be asking it to prove Euler’s theorems! It confuses the AI model and can lead to whacky stuff if the AI misunderstands the context. Again, no harm in starting a brand new channel for your needs.
With all of this in mind, and when I practice these tips, I find myself in a different position on the scale of AI users: neither addicted nor a stay-away, I’m a STOW-away. I use AI as my secret weapon when necessary, but do all the corrections and overseeing myself. AI is always there to help me, no matter what, and as I use it freely to assist me to be the best student I can possibly be.
How can YOU become the AI Grandmaster?
So what are my experiences with AI and how has it helped me? Honestly, it’s become a pretty integral part of my brainstorming process.
EssayGPT
For example, for English essays, we have to compile the quotes, analysis, and the thesis in a template. In general, AI helps me make my writing more cohesive and professional, and asking it to grade my essay according to an entered rubric is quite helpful. Another major use point for me was preparation for my English final on Shakespearean sonnets, where ChatGPT and other models helped me not only understand Shakespeare and the elements of poetry but also were able to find Shakespearean sonnet for me to practice without me having to go to online libraries and manually find them (Fun fact: one of the sonnets GPT gave me to practice actually came on the final exam!). Also, in particularly difficult passages or when I’m having a brain fart, AI helps steer me on the right path by providing specific suggestions that I use as a launchpad for my own ideas.
Organizing my AIs
I also use multiple channels on AI platforms for different subjects: I have a channel for Chemistry, History, Math, English, etc. which helps the AI to adjust to my questions. I can safely put all of my chem homework and coursework questions in the first channel, so when exam time comes around, I can just ask the AI to give me practice questions based on previous discussions. It already knows my level and knows what areas to target without needing me to upload/copypaste a bunch of material.
One of my favorite features (although a bit wonky right now) is the “upload” option, like ChatGPT has. I remembered on my World History final last year, I just uploaded a pdf of the study guide and slides and just asked it to give me multiple-choice questions for studying. AI practice tests help build active recall study techniques; and so while my friends were filling out the study guide from their notes, I was getting hard MCQs from AIs that had the study material. You better believe I got an A on that exam!
As the ultimate librarian
I don’t just use AI for academic purposes. I find it very useful for my debate research as well (post coming soon!). Instead of using AI as a virtual teacher, I use it as an intelligent librarian with the library for the whole internet at its disposal. For example, the January Public Forum debate topic is whether the African Union should recognize Somaliland as a sovereign nation. Before this, I had very little background knowledge and couldn’t begin formulating a case until I had some. Instead of getting lost on google in sketchy articles, I just copypasted the debate topic and brief to AI and let it develop a comprehensive review for me depending on current events: I learnt about the Ethiopia-Egypt scandal, the MoU between Ethiopia and Somaliland, and Somalian pirates! Indeed, after I’ve formed my case as well, I can ask it to judge it based on whether the judge is a layman or an experienced debater—this is a much more sustainable, near infinite source to bounce ideas off of then people like parents, friends, or siblings as they might be busy.
Tool, Drug, and Helicopter
All in all, I’ve come to appreciate AI as a useful tool to help me become the best student I can possibly be. It isn’t a friend or an aide: it’s a tool, a very effective tool, that when used correctly, can quadruple your efficiency and accuracy. Although pessimists would cite the pitfalls of overreliancy in using AI, I would argue that practices like asking AI for practice tests and helping understanding material builds self-reliance and helps students adjust to the nature of having to understand things themselves that makes them more confident in their capabilities.
Initially, AI was like a new drug: there were the addicts and misusers versus the anti-vaxxers. I don’t think either is right: I want to use it as a MEDICINAL drug. use it when it’s helpful, helpful to YOU. However, it’s not a whole new person: it simply keeps you going and speeds up the recovery process. You can take it as many times as you want, but in the end, it’s what you do in life that influence how you use it, not IT destroying YOUR life.
Think of it this way: imagine your task (assignment, project, exam) is likened to going from point A to point B. Traditional work would be represented by a squiggly, handwritten line that goes all over the page before reaching point B. Using AI is like having a helicopter that flies you in the direction of B. Because AI isn’t perfect (yet…), you will still have to do your own work: but better to use a helicopter to cross the Rockies then walk on foot.
Now, a hard task doesn’t give me that sinking feeling in my stomach: instead, the excitement rushes through me as I know that I’ve securitized myself using AI. Forget quality over quantity: here comes quality AND quantity!
While juggling the question of what role AI plays in my student life, I’ve come to many conclusions, and I’m still not sure what it is: all I know is that it has helped me tremendously.
One thing we can’t ChatGPT: The Future
But we can only speculate what this technological marvel will bring in the next 10 years. The true potential of AI has yet to be seen. It’s a developing field and whoever can unlock it’s power will be very successful with it. For now though, remember to use it intelligently and not allow it to damage your learning. Ask it questions, upload files, and play around! There’s no harm in exploring and pushing the boundaries of AI. Heck, I enjoy messing with its “no political answers” OpenAI tried to lock it with (check out youtube videos that do this: funny ones include videos where someone tries to outsmart it/concede the existence of god).
One thing that both sides of the bell curve can agree with regarding AI can be summarized in a funny Hindi proverb: नकल् में अकल्; “nuh-kul mein uh-kul”; “intelligence in copying”
Perhaps the new lunar year shouldn’t be the Year of the Snake: Rather, 2025 shall be the Year of Artificial Intelligence.
Roko’s Basilisk anyone? Like and Sub!