Schools cannot build classrooms that block out artificial intelligence. Instead, educators need to redesign assessments with the assumption that AI tools are present and accessible to students.
The premise is straightforward: attempts to eliminate AI from learning environments fail. Students already use generative tools to brainstorm, summarize, translate, and draft. Pretending otherwise leaves educators unprepared to evaluate authentic student thinking.
This shift requires fundamentally different assessment design. Rather than prohibiting AI use, educators should create tasks that demand human judgment, synthesis, and reasoning that AI alone cannot produce. Open-ended problems requiring students to apply concepts to novel situations, defend positions with evidence, and explain their thinking become harder to complete through AI shortcuts alone.
Schools implementing this approach redesign major assessments around complex, real-world challenges. A history class might ask students to analyze primary sources and construct arguments about causation, rather than asking them to memorize facts that AI can instantly retrieve. A math class might focus on problem-solving processes and mathematical reasoning instead of computational accuracy, which AI handles trivially.
The practical implication shifts assessment from product to process. Teachers increasingly ask students to show their work, explain their reasoning, document their thinking, and reflect on their decisions. Oral presentations, Socratic seminars, and interactive conferences become more common because they capture thinking in real time.
This approach also requires honesty about when and how students use AI. Some educators now ask students to disclose their AI use on assignments, similar to citing human sources. Others build AI-assisted work into curricula deliberately, teaching students to use these tools effectively while maintaining academic integrity.
The goal is not to eliminate AI from the classroom. The goal is to ensure that human thinking remains central. Educators who design assessments around what humans do best—interpret context, make judgments under uncertainty, create meaning from information—build systems that work with AI rather than against it.