From Certificate to Contract: Private AI certification, supervised internship and employment in the Japan-Cambodia skills corridor

This paper asks whether a private, non-accredited AI certification programme, bundled with a supervised internship and an employer network, can move young Cambodians into formal employment faster than the alternatives available to them, and whether the financing that makes it accessible does more good than harm. The programme teaches in nine languages through AI tutors, registers credentials on a public blockchain, charges US$30 per lesson, and connects graduates to member firms that host internships and hire from the pool. Three mechanisms could explain any effect: human capital accumulation, signalling where credential information is thin, and network-based matching. The paper sets out the corridor model, subjects the operator's own operating figures to arithmetic and finds that they reconcile only on one reading, in which roughly 87 percent of learners have never paid; states the six conditions under which the model could work; specifies the findings that would disprove it; and proposes a measurement protocol an independent researcher could run without the author's involvement. It is written by a practitioner about an entity she leads. The conflict is disclosed on every page and constrains how far the conclusions can be pushed. Declared interests. The author is founder and chairwoman of E.M.B.O.A Co., Ltd. (Cambodia), a member of E.M.B.O.A LLC, and a proposed founding member of Emboa Japan GK. Global College of Technology & Innovation is an entity she leads. This is a working paper, not a peer-reviewed article.