Student and teacher concerns about AI’s effects on critical thinking are intensifying, alongside evidence that usage for homework is rising. A RAND survey found that nearly 7 in 10 middle, high school, and college students reported concern that AI use erodes critical thinking skills, with concern rising sharply between May and December 2025. Usage tracked closely to that worry: from May to December 2025, reported use of AI for homework increased across grade levels, including a jump among middle school students and high school students. The survey also found that generative AI chatbots were the most used tools, suggesting that broad “chat” systems—not narrowly supervised tutoring platforms—are driving student exposure. The report does not prove harm, but it documents a credibility and dependency issue: students are increasingly using AI while simultaneously believing it weakens independent thinking. That mismatch is likely to feed institutional responses—policy updates, assessment redesign, and training for both instructors and students. For higher ed decision-makers, the takeaway is operational: learning support models that rely on AI may need stronger guardrails, stronger transparency, and more measurement of learning outcomes beyond adherence to rules.