In June 2026, the Hong Kong Curriculum Development Council issued the Blueprint for Digital Education Development in Primary and Secondary Schools, which mandated all in-service teachers to undergo at least 30 hours of digital education training within each three-year cycle of 150 hours of continuing professional development. This article reviews the policy context, evidence of implementation, and ongoing tensions associated with this mandate. The analysis draws on government circulars, legislative papers, teacher surveys and international comparative frameworks to show a significant gap between the policy intent and the reality of implementation. More than 90% of teachers in Hong Kong are already using AI tools, but they rate their own proficiency at only 5 out of 10, and nearly 60% say AI adoption is driven by individual exploration in isolation, rather than systemic support at the school level. The training content still concentrates on the basic use of tools, with over 70 percent of teachers reporting that they do not use advanced artificial intelligence tools like AI agents or Retrieval Augmented Generation. However, the article proposes that the 30-hour requirement will work only if three conditions are met: if training content shifts from generic tool operation to discipline-embedded pedagogical application; if assessment frameworks shift from time-based compliance to competency based certification; and if support structures shift from fragmented individual effort to coordinated institutional ecosystems. A research agenda is proposed to investigate these conditions using longitudinal, comparative and mixed methods designs.