“I put a paper I wrote in high school through an AI detector; it came back 65 percent AI generated,” my cousin told me a few weeks ago on a phone call. Since she had gone to high school in a pre-generative AI world, this was pretty unsettling.

As AI continues to develop, it is getting harder to discern authentic writing—a challenge for both students and teachers. This is not only a challenge, but an obstacle, as my cousin experienced, where technological systems and AI itself fail to identify 100 percent AI-free writing. It had us spiraling. What if a professor wrongly believed I used AI in a paper I truly wrote myself? What if a preventive AI system flags my authentic writing for AI? At the end of the day, whose claim is more credible, mine or AI’s? With no answers, our conversation ended like most AI-related conversations do: with uncertainty.

Unsurprisingly, I have a mostly negative stance on AI, but I acknowledge that its inevitable, ever-growing presence and significance is not something that should be avoided, but rather something we should learn how to adapt to. I did not strongly feel the presence of AI in classrooms until my junior year abroad through a program at Doshisha University. My abroad program embraced AI, encouraging us to use it as a tool to support our language learning.

Unlike other translating tools, Doshisha University used an AI program that focused on “human-sounding AI translation” to translate Japanese into more natural, comprehensible English and vice versa. The AI software also offered multiple translation options for both natural English and Japanese, depending on the target language. Translation is, in many ways, a process of individual interpretation. Using this inherently unnatural AI program, ironically, felt natural, since my study abroad program encouraged all students to use AI software for grammar checks on assignments. Coming back to Bowdoin, AI use no longer feels taboo, as it did in previous years.

Upon returning to Bowdoin the following year, it was noticeable that syllabi handed out by professors often included a sentence or two discouraging or outright banning AI use on assignments. In one of my first classes back on campus, the professor required students to read an AI-generated version of their paper and reflect on the differences in the writing process across various AI tools. Beyond that, other professors are starting to use separate tools that monitor student writing as a preventative measure against AI, with other courses fundamentally changing their evaluation style to exams or oral tests.

After noticing this stark difference, I reached back out to my friend, Saya, who studies in Kyoto to ask more about AI use in her classrooms.

“Just the other day, our English homework was to use AI to polish our own essays, and I feel like more and more professors are recommending AI,” Saya said.

Reflecting on these approaches to AI in classrooms in two countries, I struggle to decide which one is inherently more effective. Although it was nice to be encouraged to use AI in my language-focused class, is the same loose approach truly beneficial across all disciplines?