Currently, there are no standard artificial intelligence usage policies throughout PUSD or within Westview, which leaves many teachers to independently decide the extent to which their students can and should utilize AI, specifically large language models (LLMs) such as Open AI’s Chat-GPT. Acknowledging this, PUSD has launched a new aspect of the Digital Citizenship Initiative that attempts to clarify for students and staff how to engineer prompts for LLM systems that aid education in an ethical and responsible way.
The initiative marks a movement toward encouraging student LLM AI usage. It is understandable that PUSD is trying to establish guidelines in this area as AI use is rapidly growing more common. However, as they do this, we earnestly hope that they keep at the forefront of their guidance that the use of LLMs has been shown to pose a risk to the critical thinking skills of young people, since it is precisely those critical thinking skills that students are here to develop. These skills are the ones that will guide our success in higher education, the workplace, and aid us in navigating life. Thus, PUSD must not allow students’ cognitive development to fall victim to the allure of any kind of LLM usage in school.
One study by MIT research scientists found that LLM users who integrated information generated by AI into their essays displayed weakened connections in the temporal and frontal lobes during the study, reflecting a bypass of deep memory encoding. In contrast, the participants who conducted research on a search engine or relied solely on their brains did not display the same impairments. The brain-only group displayed the strongest frontal lobe activity as measured by an EEG, thus committing more information to memory and benefiting more from writing the essay.
This deep memory consolidation is vital for information recall—a crucial step to the learning process and a foundational part of critical thinking.
Additionally, the environmental devastation caused by A.I. cannot be underestimated. The MIT study revealed that LLMs consume around 10 times more energy than a search engine such as Google or Safari. An article in the Harvard Business Review by a UC Riverside professor and Caltech professor found that the training process of LLMs emits hundreds of tons of carbon, consumes thousands of megawatt hours of electricity, and exacerbates limited freshwater sources by draining high quantities of freshwater evaporation for data center heat rejection.
As general LLM usage continues to grow, it’s crucial that PUSD policies are made with AI’s cognitive and environmental impact in consideration. There is little doubt that AI is here to stay and that it will play a prominent role in our adult lives. However, that is no reason for PUSD to rush into implementing its usage in any way today when the risks to young minds are too great. Decisions that hinder students’ learning and encourage unsustainable practices will not benefit PUSD students nearly as much as a focus on original thought and critical thinking.
