You fear that when your teacher reads your AI-generated essay, sirens will go off. Rather, sirens should sound throughout our school to alert us to our unpreparedness for the AI revolution. To nullify AI-assisted cheating, help students develop AI-era skills, promote deeper human connections, and more effectively teach critical thinking, Milton should allow AI on all take-home assignments.
In The Milton Paper’s 2026 State of the Acad, 41% of students admitted to cheating with AI. Moreover, among the 67 student respondents to a survey, 37% rated their stress level at being falsely accused of cheating with AI as a four or five out of five. In response to this brutal reality, English Teacher Malinda Polk wrote that she “hate[s]” the need to verify student work in this landscape, where natural typing simulators and humanizers seek to bypass AI-writing detectors.
Currently, Milton’s faculty, either individually or by departmental policy, have full control over their class’s rules about AI. This state of affairs broadly results in an education system that does not prepare students for the AI era, in which AI can perform tasks that are impractical for humans. For instance, Claude Cowork can do various finance, marketing, and legal tasks autonomously and thereby caused a trillion-dollar sell-off in the stocks of traditional software services, a selling that signifies a shift in human work toward the management of autonomous AI agents. Possessing classical critical thinking skills but limited experience with AI, Miltonians will fail to pioneer in the rapidly evolving world as they have in the past, according to Computer Science Department Chair Chris Hales. Thus, when students were asked the extent they felt Milton prepared them for the AI era, the median rating was one out of five.
The median teacher reported using AI for only ten minutes per week, and 67% of teachers admitted to giving zero assignments that permitted AI use. If Milton allowed AI on all take-home assignments, faculty and students could learn about the capabilities of AI tools when AI succeeds or fails to automate the assignment. Building from this knowledge, teachers could then assign projects that require human-AI collaboration. For example, Aarav Agrawal ’27 described how making a video game based on the novel 1984 for a Class II English project led him to think more critically about his use of AI.
In their statement asking freshmen to join teachers in keeping Class IV English AI-free, the English department argues that efficiency gains from AI detract from collective discovery around the Harkness table. Students lose this experience when they ask a chatbot for an answer, the document says. However, a chatbot could challenge students’ ideas and thereby make them more robust. Students could also sharpen their critical thinking skills by evaluating the chatbot’s advantages and limitations. Then, students could present their more developed ideas to the table to prompt deeper discussion. According to The New York Times columnist Ezra Klein, improvements in AI have led him to have deeper conversations with his doctor, therapist, research team, and video editor.
Full permission to use AI on take-home assignments could promote critical thinking: students could develop their thoughts by collaborating with, and also by evaluating the contributions of, AI. Whereas AI can quickly analyze large volumes of information, humans can better evoke emotion. The differences between AI and human capabilities mean that human-AI collaboration often yields better results: even if underclassmen have less-developed writing skills, their lived experiences allow them to evoke stronger feelings. To help students discover their unique voice, teachers could simply require that essays pass AI-writing detectors.
Most foundational knowledge is taught through in-class essays and exams. By testing and interpreting research separately, science teachers could break larger skills, such as writing a lab report, into testable chunks. Requiring students to start drafts in class without access to AI would ensure they maintain the ability to brainstorm.
Milton should prioritize teaching the most valuable skills within the limited time available to educate students. Thus, we should focus on AI literacy and human strengths, such as developing deep relationships, rather than on skills that AI can replace. History classes prioritize source analysis over memorizing the Dewey Decimal System. Since AI can sift through hundreds of sources in minutes, whereas humans take hours to find relevant quotes, history teachers should allow AI research tools to help students find sources. While students might not randomly find an interesting quote, they gain a much more important ability: deep analytical thinking.
That said, critics may caution against AI’s environmental impact. AI datacenters, however, have a minimal ecological footprint. Datacenters account for roughly 0.51% of global carbon dioxide emissions from their consumption of electricity, per a 2025 report by the International Energy Agency. According to a Google Cloud paper, the median text prompt on the Gemini app emits an equivalent of only 0.03 grams of carbon dioxide and evaporates only 0.26 milliliters of water back into the water cycle.
In addition AI labs have indeed committed various potential ethical violations: training on copyrighted materials, developing biased systems, underpaying low-wage laborers to classify disturbing AI-generated content, widening wealth inequality, and failing to safeguard models to protect users’ mental health sufficiently. Precisely because of these ethical violations, we must learn about AI. This learning requires practical experience. Rather than banning AI, teachers must assign tasks that prompt discussion of ethical AI use.
We should apply the same standard to AI and computers. Milton requires Upper School students to have a laptop for classes, even though the cobalt used in computer batteries comes from human rights violations: per Amnesty International, cobalt mines often use child labor and fail to protect workers properly.
To ensure equal access to AI, Milton should provide a $20-per-month grant to all students for a premium AI subscription. Flint AI, which the school currently purchases, lacks essential autonomous-agent capabilities. The benefits of experience using AI far outweigh the monetary and ethical costs of the payment.
Therefore, to nullify AI-assisted cheating, help students develop AI literacy, facilitate deeper connections, and strengthen students’ critical thinking and foundational knowledge, Milton should allow AI on all take-home assignments. Let us begin a new period of discovery, in which teachers and students collaborate to explore the human in the AI era.