The Last Assignment I Couldn’t Outsource

Getting home late from a rain-delayed baseball game sucks. Whenever this happens, all I want is to take my damp pants off, blast the air conditioning, shower, and go to sleep. But high school doesn’t let me do that. No, on this particular night, it sent me to Schoology, where I found a thirty-minute Spanish listening exercise followed by a four-paragraph written response. At the top of the page was a note: “Please note that this assignment is not simply checked for completion; anyone who fails to produce a response that meets rubric expectations will lose exemption privileges.”

I absolutely did not want to do that assignment. My study’s air conditioner had stopped working, and I had back-to-back biology and multivariable calculus exams early the next morning. The last thing I needed to do was listen to a Chilean newscaster ramble on about avalanche risk mitigation in rural Andean communities. But I was on thin ice in Spanish. Senioritis had left my previous quarter’s formatives at a collection of C+, D+, and Fs. Losing exemption privilege meant that those exams I had strategically ignored would come back to hurt. This assignment had to be done. 

I knew in the back of my head that this could have been done easily with ChatGPT. It would give me a perfect response with flawless grammar and a train of thought I could not produce even in English. I’ve seen it plenty of times in my friends’ work. 

Sure, I know all the downsides of using AI for schoolwork. Teachers constantly talk about how it damages learning and weakens critical thinking. But sitting there exhausted after midnight, none of those arguments felt relevant. What mattered was that AI offered a direct path to success in the exact system I had spent the last four years learning how to navigate.

For others, defining success in life is difficult: they need to establish a clear goal (make a million dollars? get drafted by an NBA team? lose 5 kg?) and break it down into actionable steps. But we high school students have it easy. At this stage in life, our clear goal is to get into the best universities that we can. To do so, we need to excel academically, and SAS makes excellence measurable: From the smallest formative assignments to the grade-defining final projects, every task comes with a rubric that spells out exactly what “exemplary” looks like. With that kind of clarity, outsourcing success to a machine is incredibly easy. All you have to do is paste the task sheet into the large language model (LLM) of your choice and instruct it to “strictly ensure that all items listed in the exemplary column are achieved and easily identifiable.” The machine does not care how tired you are, how badly you procrastinated, or whether you actually understand the material; it simply optimises for the rubric better than most students can. In a system where success is reduced to measurable deliverables, LLMs designed to optimise clear deliverables make the thinking and effort behind academic excellence feel unnecessary. Why should I spend hours writing in my fourth language when a free machine can write an article on Andean avalanche infrastructure with more detail than a Chilean newscaster will ever have in his brief?

ChatGPT and Claude have worked flawlessly for countless high school students. But as we graduate, we face a daunting reality: the system that enables us to outsource success to artificial intelligence is about to disappear.

After I (hopefully) graduate from my dream university, I will no longer have clear goals on what I want to achieve, let alone rubrics that dictate whether I will be approaching expectations, meeting them, or exceeding them. The real world does not lob softball Spanish audio exercises or standards-based rhetorical analysis essays on the works of F. Scott Fitzgerald. Out there, success will become much harder because I’ll have to define my own goals. Maybe success means building a meaningful career, forming deep relationships, or even just staying healthy. But whatever the goal is, there won’t be predetermined answers hidden somewhere in some list for an LLM to decode. Getting to any of those goals will take un-outsourceable skills: experience, grit, creativity, judgement (both critical and ethical), and God forbid, emotional intelligence.

So, as I graduate, I must bid farewell to a system where success is clear and therefore fully outsourcable. I won’t pretend to miss a world where all that sat between me and an A was a well-crafted prompt. But I cannot bring it with me, and I cannot outsource my way through whatever comes next. The version of success waiting for me on the other side doesn’t come with an exemplary column. It doesn’t come with a column at all. 

Which brings me back to the Spanish assignment. 

I refused to use AI for it, but not because I suddenly transformed into some noble defender of academic integrity. Frankly, the smarter short-term decision would’ve been to generate an essay and go to sleep. But I realised that if I start outsourcing thinking and comprehension now, I would eventually lose all ability to depend on myself to get things done, even if those things could not be outsourced.

So instead, I shut my laptop, accepted the zero, and received a stern email home informing my parents that I was no longer eligible for exemptions.

Very well.

The night before the test, I sat down and listened to the audio properly. It was frustrating, given that Chilean Spanish sounds to learners the way Scottish English sounds to foreign ears. 

But maybe it was worth it. After a few hours, I actually felt like I learned something. And I did pass the test the next day. Maybe that means success is possible without outsourcing. The process sucks, but it’ll be good to know for the future, where there are no more task sheets to make the perfect prompt, and where there are certainly no exemptions.