We’re living in a world flooded by Artificial Intelligence (AI). Just recently, OpenAI released Sora 2, their improved video generation model for public use, and the results look extremely promising. Whereas it once was easy to discern reality from generated content, it is now significantly harder.

So, it’s safe to say that AI is not going away anytime soon. According to a study Campus Technology conducted in 2024, 86% of students utilize AI over the course of their academic career. Because of this, schools are starting to emphasize AI literacy. 

At Taipei American School (TAS), we’ve been taught that AI can be a powerful learning tool if properly utilized. “I use it to study or get additional information on something … [and] when I have a quiz, … I’ll use AI to help quiz me,” Ira Chavan (‘29) explains. Still, Ira admits that she relies on it too much, making her “a bit lazy in [her] thought process”.

Here, Ira points out a deadly flaw in our usage of the technology—in the process of learning to use AI, we’ve deluded ourselves into believing that we’re only giving it tasks we could’ve easily completed ourselves. Take brainstorming: In the grand scheme of things, brainstorming seems like an insignificant task. However, once we begin to justify AI use with this line of reasoning, it becomes a slippery slope; we start using it to complete far larger tasks, so long as we feel as if we are producing the final product ourselves. 

This makes it concerningly easy for us to fall prey to becoming over-reliant on AI. This happens when we begin to consistently utilize and rely on generative AI models to complete tasks without thinking through the output ourselves. By doing this, we’re reducing our own time spent critically evaluating the output or subject ourselves. We stop seeing AI as a tool, but as a magical, omnipotent technology that can instantly solve all of our problems for us. 

As a study conducted by MIT suggests, this proves to be detrimental.

In the study, students were wired up to devices to monitor brain activity and asked to research and produce an essay over the course of four months. The students were split into three groups: one was asked to guide ChatGPT to write the essay for them, one was allowed to use search engines, and the last was only permitted to conduct hands-on research. The study found that the ChatGPT-only group had the lowest level of brain activity, and 83% couldn’t even recall key points in their own essays. Perhaps even more concerning is how the ChatGPT-only group’s brain activity lessened even after they stopped using AI. Ultimately, the study suggests that over-reliance on the outputs from generative AI can eventually result in faster cognitive decline, weaker brain connectivity, and poorer memory retention.

Of course, the way forward isn’t to stop using AI; whether we like it or not, it’s going to be an integral part of our future. Thus, it is imperative that schools make an effort to inform students of the risks involved beyond the surface level. Rules won’t hold if people believe it to be backed by weak evidence. Currently, TAS seems to expect all students to know the harms: students who use AI to complete their work won’t learn the material. And therein lies the problem—that’s the extent of most students’ AI literacy. Unfortunately, lack of learning most certainly isn’t the only harm. We’re on the Titanic, and what most see is only the tip of the iceberg. While it is unreasonable to educate students on every possible risk, an effort should at least be made.

There’s no such thing as a free lunch, and AI is definitely not free: it has the potential to be a massive cost to our mental ability, the one thing that distinguishes us from other species. The second we start to use it as a shortcut instead of as a tool is when we become victims of AI. We should be the ones wielding it, not the other way around.

If there’s any time to be cautious, it’s now. The AI wave is here, and if we’re not careful, we’re going to get dragged under.