Now that artificial intelligence is everywhere, from phones and computers to navigation systems, does it also belong in the classroom? It might be tempting for students to copy and paste their essay prompt into ChatGPT and have it write their paper, but that’s not effective learning. Instead, Baldwin faculty aim to regulate student AI use to be healthy and only used when necessary.
According to Mrs. Christie Reed, Head of Upper School, AI can be a tool to help students, but it should never become the only resource. Some early studies suggest that the frequent use of AI and similar technologies may affect memory and overall cognitive abilities.
“You lose the ability to think critically,” Mrs. Reed said. “Students need to learn how to evaluate AI writing, because while AI may replace many things, I hope it doesn’t replace fundamental critical thinking.”
Teachers can often recognize AI-generated writing by identifying hallucinations. Hallucinations occur when a program responds to a prompt, but produces output not based on accurate information. It may sound correct and knowledgeable, but it is not grounded in true facts.
“That’s a dead giveaway that text is AI-generated or assisted by AI,” Mrs. Reed said.
Some other signs that text is AI-generated or AI assisted is when students change their style of writing, using words they have never used before or writing in a more professional style.
Aside from being able to identify AI-generated work, Upper School faculty are exploring how students and teachers can integrate AI into the classroom. This way, it can support creativity and critical thought rather than replace them.
“AI is a tool we should use, but it’s important not to abuse it,” Dr. James Heverin, Director of Technology and teacher of AI-related courses, said. “Be mindful of what you are using it for.”
In Dr. Heverin’s Artificial Intelligence and Society course, students used AI to create recipes, art, and songs. Then, they recreated the same products without using AI, asking teachers to come in and try to guess which versions were AI-generated. According to Dr. Heverin, the results were about 50-50. Despite some clear indicators of AI use, it has advanced so much that it is becoming harder to recognize.
AI’s growing undetectability has inspired the creation of AI detectors. However, both Mrs. Reed and Dr. Heverin believe that AI detectors are unreliable.
At The Baldwin School, AI detectors are rarely used. Instead, the English and History departments have integrated the software Rumi into writing assignments, which gives teachers insight into each student’s individual writing style. For instance, Rumi tracks every keystroke as well as detecting if text has been copy and pasted. After submitting your writing, the teacher gets an analysis of students’ writing.
Although detecting AI is challenging, that does not mean that Baldwin students should be using it more often. Baldwin teachers have integrated learning tools into their classrooms that prevents the unnecessary use of it.
Mrs. Reed and Dr. Heverin also emphasize that understanding teacher expectations is crucial. Before students use AI, they should make sure to ask teachers in what ways they are allowed to use it. For example, students may be allowed to use it to make a study guide for a test, but not to write an outline for their essay.
Overall, Baldwin teachers and administration recognize that it is impossible to completely eliminate AI use, given how pervasive it has become. Instead, the goal is to teach students how to use it responsibly.
As Dr. Heverin said, “AI can be beneficial in some ways, but don’t rely on it.”
