ChatGPT in Education: The Debate Continues – New York City Schools Ban AI for Student Usezinkerz
Each year, around the end of December, children all around the world gather together with their families for the most anticipated event of their years: Kurtis Conner’s annual Christmas movie review video. For those who do not celebrate this event or are unaware, Kurtis Conner is Canada’s most-mustached man, the mayor of Kurtistown, and a YouTuber who has gained popularity through his comedic edits and tackling of social topics through the lens of goofiness. His most popular videos are of him reviewing the worst Christmas movies that Hulu, Netflix, and Hallmark have to offer.
But, 2022 was not a typical year – not for Buckingham Palace or Kurtis Conner. Unlike Lizzy, Conner thought outside of the box and decided to create his own Christmas movies…with a twist. Using Runway, a program that is able to train artificial intelligence to perform specific tasks, Conner used the scripts to Die Hard, Santa Clause, Home Alone, and Bad Santa to train OpenAI and ChatGPT bots and teach them to write scenes for Christmas movies. The purpose of this was to test the AI’s capabilities and to commentate on the scenes, which he filmed himself, for YouTube. During his video, Kurtis mentions that the process still took him hours as it included uploading entire scripts, and he still needed to edit the scripts from ChatGPT to make them more legible. The video contains adult language, so we will not include a link here, but we also can’t stop you if you go to his YouTube channel, click on his most recent video (as of January 9, 2023), and skimming to the 10:55 mark to watch the scenes that were written by AI bots. If you do, you’ll come to a similar conclusion that we arrived at: these are nonsensical, and they stink! Despite learning the entire script of the movie they are mimicking, the ChatGPT Ai is unable to formulate cohesive characters, confuses about which ones are supposed to be speaking and participating in a scene, and totally mishandles the English language.
In his introduction, Conner mentions that ChatGPT has been used to write emails in corporate correspondence, papers in academic atmospheres, and blog posts for tableted tabloids. You can feel confident in knowing that this blog was written by a human: robots aren’t capable of the kind of alliteration that I just pulled off a sentence ago. Recently, Vice posted an article that adds to Conner’s claim that ChatGPT has been used in education as well as our claims from last month and October: New York City schools are banning the use of ChatGPT for students and teachers alike.
Good job, New York!
As a lifelong Bostonian, those are four words that I never wanted to say. But, as they are warranted, I will repeat them: Good job, New York! I really liked that part of the season where you laid down and forgot how to play baseball against the Astros. But, regarding education, New York City has banned access to ChatGPT within schools and on school computers for the entirety of all of their districts for the sake of “safety and accuracy” regarding student work. New York’s education department released a statement that they were concerned with ChatGPT’s outputs diverging from the typically accepted completion of homework and that the AI has a negative impact when it comes to a student’s education.
We couldn’t agree more! When I tried to ask ChatGPT itself what it thought of the ban, I was met with this message:
Thankfully, Samantha Cole at Vice had the same idea and ChatGPT’s response was: “It is important to consider the potential risks and benefits of using ChatGPT in education, and to carefully weigh the evidence before making a decision. It is also essential to listen to the perspectives and concerns of all stakeholders, including educators, students, and parents, in order to make informed and fair decisions.” Additionally, it should be said that Shakespearean writing cannot be considered impressive by ChatGPT but by the developers. The AI is not developing these words – it is a screen programmed to appear as though the AI is writing it.
It’s essentially a GIF that pops up to say “ChatGPT machine broken.”
It would be ridiculous for me to argue with this statement as it is two sentences from a robot that can’t think, only reproduces old thoughts from the internet. The benefits of ChatGPT are that sentences will be written and a student can hand them in to make an attempt at completing an assignment. I am going to assume the perspectives of each of the groups ChatGPT listed:
Educators – Passing in work that is not a student’s own is cheating and detrimental to their individual education.
Students – Awesome! Now, I have more time to stick peanut butter to the roof of the dog’s mouth and watch them try to lick it off!
Parents (mothers) – I want my child to learn! ChatGPT doesn’t help them do that!
(fathers) – Sorry, I am busy growing my mustache to look like Kurtis Conner.
Can I get back to you on that?
The purpose of writing assignments is to test a student’s ability to comprehend the assigned work, to come up with their own arguments, and to coherently express the ideas and emotions that they have in their minds. An argument for ChatGPT could be to use calculators as an example. Even as a mid-20’s writer, I remember teachers saying that we needed to learn to do math without using calculators because, in the world of adults, we weren’t going to be able to carry calculators everywhere. We couldn’t rely on those machines because we wouldn’t always have them. Unfortunately, Mrs. Hobart (my second grade teacher) didn’t have the future-telling abilities to predict smartphones and their capabilities for also being calculators. Personally, I still use lattice multiplication, and I have not done math since I changed from a chemistry major to a philosophy major in 2018. However, writing is an entirely different realm of thinking.
Any subject, from English and psychology to chemistry and even math, the ability to reason through a problem and to find a solution is going to be a significant skill to have. Carpenters, teachers, financial advisors, Chipotle cooks, and college counselors all need to have the ability to communicate what is in their minds and to finesse solutions in their decisions. Using ChatGPT dissolves the lessons that students are meant to learn regarding communicative skills. Relying on a robot to type papers and blogs reduces a student’s future ability to choose an apartment, explain to a judge why they thought the speed limit was 50 and not 30, and write a cover letter explaining why they want and are a good candidate for a job. For the introverts out there, I’m sorry- it’s a social world and every industry is reliant on communication and reasoning to conduct business. Calculators were not always available, but now they are. Thinking and reasoning have always been available, but writing and completing school work with honesty is how students learn to use their minds. ChatGPT hinders this growth.
If you’re a student or know a student who would like additional help writing in their classes or in preparation for their college educations, get a free consultation with Zinkerz here! We will be able to create a unique study plan that best helps a student develop expertise in their subjects and can connect the student to teachers and counselors that are able to masterfully guide them throughout the entirety of their high school careers!
For teachers and parents: GLTR. GLTR, which stands for Giant Language model Test Room, is a forensic website that helps to determine whether a student or an AI bot wrote a piece. The way GLTR works is this: you enter a piece into the text box. This text is then analyzed by GLTR using entropy. This means that it reads the text backwards; it looks at each word in the context of the piece and asks itself “what are the 10 most popular words that would precede this word to make sense in this scenario.” If it’s in the top 10 most logical words, the word is highlighted green. The word is highlighted yellow if it’s top 100, red for top 1000, and violet if it’s outside these realms. So, if the entire or most of the piece is in green, that means it was most likely generated by an AI bot. Much like Alan Turing in The Imitation Game, AI like ChatGPT works off of logic. They do not write what “feels” or “looks” good. They write what makes the most sense in answering the prompt. In doing so, each word and letter is given its most logical place, according to the robot, without any emotion and GLTR has cracked their enigma code for cheating. All of this is under the assumption that the AI produces legible scripts, unlike in the Kurtis Conner video.
When it comes to apartment rent prices and in banning ChatGPT from schools, we here in Boston are fine with finishing second to New York City…so long as the Red Sox finish above the Yankees.
Leave a Reply