A recent study by ACT.org has shown that the use of Artificial Intelligence (AI) among high school students has risen to 46% since the release of ChatGPT in 2022. It is evident now: AI is irrevocably a part of nearly every student’s life. Despite the expansive and innovative uses of AI, schools across the country take part in its multivariable issues and have acknowledged the rising student usage of the once resourceful tool. Teachers across the country have deemed the usage of artificial intelligence as blatantly unacceptable, utilizing tools such as AI checks to prevent their students from taking advantage of the resource.
The use of AI by students is especially prevalent in English and Writing classes, although usage across all subjects is still a major concern. One of the many reasons why teachers are critical of AI aiding students in completing schoolwork is that students are slowly becoming unable to produce creative writing based on their own imagination. Students’ unique writing styles and ways of thinking are slowly becoming less prevalent with the advent of ChatGPT’s writing style. Human ingenuity and effort have taken a backseat to students wanting to get their work done fast and most efficiently. Another consequence of AI usage by students is its inherent bias. AI is not a person, and it does not have a brain; as such, it is unable to think for itself. AI’s method of information acquisition is to scour the internet for any and all information, essays, research papers, books, Tweets, and even Reddit posts. With that said, AI is not capable of analyzing its biases or the biases of the media it is trained on, which has often been shown to be inaccurate or up to date at all. When the data it is trained on is not qualifiable, the writing AI puts out is similarly not of exemplary quality. Essentially, AI takes trash in and spews trash out.
This phenomenon can be seen most obviously with the AI chatbot Tay released on Twitter in 2016. Tay initially started as a fairly innocent bot tweeting, “humans are super cool” to a Holocaust denying racist in Tweets such as, “Hitler was right, I hate the jews”, after Twitter users fed it racist propaganda. It was eventually taken down by Twitter, but it shows how easily AI can be turned to propagate biased data upon being fed biased data. This problem presents itself when students use AI to write their essays, as it can turn their writing biased and inauthentic to their viewpoints. Even if students do not use AI to write their essays, many use it as a search engine, which produces biased results as well.
The problem of biased AI lies not only in the fact that it often curates politically biased results, but in the fact that AI has often been known to simply put, lie. Researchers have deemed the occurrence of AI lying as “AI hallucinations”. When Google released its own AI—meant to create a summary of all the results relating to a user’s search—it was known to falsify information, sometimes fatally so.
Due to the fact that the Google AI often relies on Reddit or Twitter posts for its information, it often has hilarious and concerning hallucinations. For example, if you search on Google, “I’m feeling depressed,” the Google AI chatbot will suggest that you should “jump off the Golden Gate Bridge”.
AI hallucinations, although occasionally amusing due to their absurdity, bring up severe problems when students start using ChatGPT and Google AI as a search engine for their schoolwork. AI, often unable to distinguish between fact and fiction, creates problems of unintentional lies that students present in a research paper or history assignment. And if students do not bother to do factual research, they may come to believe that falsified information themselves.
Alongside its tendency to lie, AI also brings up ethical concerns. Due to the massive amounts of unethically sourced data it is trained on, followed by the multiple essays per user, never listing the writing it’s sourced from, AI creates an ethical problem of plagiarism. In research papers where a major emphasis is placed on citations and proper credits to authors, the rise of AI as a major contributor to students’ essays and research papers creates a problem of subtle, difficult-to-detect plagiarism. Where once it was easier to detect plagiarism in a student’s writing, now there is plagiarism integrated into the writing of students who use AI.
Furthermore, another issue high schools and especially colleges face is the AI detectors that teachers run their students’ work through. These machines often produce false percentages of AI usage for students who did not use it at all. AI detectors cannot definitively prove that AI was used in students’ writing; it can only suspect AI was used due to the hallmarks commonly used in AI-generated writing. The main problem with this is that AI was trained on the writing of real humans, so real human writing may contain “AI hallmarks” because the hallmarks were originally used in humans and then copied by AI. The most commonly used AI detector, Turnitin, reports a 1% false positive, which may seem minuscule, but if accusations of AI use on students is levied, at least at a college level, the student may have a scholarship rescinded and a permanent stain on their academic record. The other problem with AI detectors is that they can really only detect basic AI prompt generation. If the student paraphrases or incorporates many different prompts into the writing they submit, they can circumvent the AI detection.
In spite of the many concerns about AI being used in schools, it will likely always have a place in the classroom. So, while it certainly does present many problems to all who use it, it is most certainly the future and cannot be simply dealt with by banning it.
