A new poll from the Lumina Foundation-Gallup State of Higher Education study finds student AI use is widespread, including when institutions discourage or prohibit it. The majority of roughly 4,000 students surveyed said they use tools such as ChatGPT, Microsoft Copilot, or Google Gemini daily or weekly for coursework, while many also reported that professors have not set clear AI policies. The data suggests a mismatch between institutional rules and student practice, with implications for classroom integrity, student support, and instructional design. Only a small share of students said their schools encourage AI use as much as possible, while a significant portion reported weekly use despite discouragement or bans. The survey also highlights a second-order effect: many students say generative AI is influencing their major or enrollment decisions. Leaders are being pushed toward “clarity” on expectations, ethical guidance, and training so AI literacy becomes part of learning design rather than an after-the-fact enforcement problem.