A classroom example from La Vista High School in California shows how generative AI can be used to create more relevant math assignments for multilingual, low-income students, but also underscores the limits of the broader “personalized learning” promise. A math teacher used a large language model to generate workforce-linked questions using U.S. Department of Labor data, helping students connect rate-of-change concepts to career outcomes. The report positions the approach as a practical tool for teachers—reducing hours spent designing lessons—rather than a full replacement for instruction. It also captures real classroom moments, including a student’s reaction to how gendered income patterns appear in the data. The wider context in the piece is skepticism driven by years of overhyped AI customization efforts. Educators and researchers are still testing whether AI-driven personalization can reliably improve understanding, engagement, and persistence. For districts considering secondary math interventions, the takeaway is that AI may help with lesson design and relevance—if teachers remain in charge of accuracy, scaffolding, and student interaction.