This article examines how different configurations of artificial intelligence (AI) in education may reshape teachers’ opportunities to experience teaching as meaningful work. Using a scenario methodology, we develop three plausible scenarios grounded in existing educational AI: Labor-Replacing Classrooms, where AI tutors displace core instructional tasks and teachers are redeployed into surveillance and exception-handling; AI-Managed Teaching, where teachers remain central but are guided and evaluated through dashboards, nudges, and performance indicators; and Human–AI Teaming, where AI is implemented as a teacher-governed, co-designed support intended to augment professional judgment. Drawing on work-design theory and scholarship on meaningful work, we assess each scenario against five dimensions: task integrity, skill cultivation and use, task significance, autonomy, and belongingness. The analysis indicates that labor substitution and the pervasive use of algorithmic surveillance threaten the integrity of pedagogical processes, potentially constraining professional autonomy and shifting the focus of teaching toward metrics and outputs that are easily processed by machines. While human–AI collaboration has the potential to enhance agency and foster collegial inquiry, it remains fragile and could lead to overdependence on AI systems if governance structures and workflows are not kept under scrutiny. Overall, the trajectory of AI in education depends on policy, on which technologies are adopted, and on broader institutional choices that can either erode or safeguard meaningful teaching.