The top of the pathway is not about using AI better. It is about teaching other manufacturing people to use it well, in a training room, at shift change, beside a line or in a college classroom, which is a different skill and a larger responsibility. An instructor who teaches a rule wrongly makes the error once for every learner, on every shift, for as long as the material stays in the binder.
The first three modules set the foundations. Module one starts with who is in the room: a thirty-year machinist, a new hire, a quality engineer and a supervisor who did not want to come, all with production to run. It explains why telling a shift the rule changes little and what the two verified studies on trust and productivity can and cannot say. Module two designs a session backwards from what a learner must do on the next shift, fitted to the time a plant will release, with three things no session ever cuts: the data screen, the named approver and the verification step. Module three keeps material accurate: law, contract, voluntary standard and certification scheme kept apart on every slide, a currency log for editions, rules, customer requirements and tools, and how to correct something already taught wrongly.
Modules four to seven are delivery. Module four facilitates the expert who argues, the silent room and the voice that takes over, runs one session on three shifts with the same accuracy, and handles the line going down mid-session. Module five runs a demonstration built around deliberate failures on fictional data: an invented citation, a withdrawn edition, a unit slip and a missed defect, each tied to its check in the debrief. Module six builds labs on fictional prints, routers, work orders and nonconformance records, and keeps real customer prints, export-controlled data, controlled unclassified information, trade secrets and employee records out of the room. Module seven teaches safely around live production: where a group may stand, the hazard assessment behind protective equipment, hazard communication, and no running machine, live robot cell or real disposition. OSHA has no AI standard, and this course qualifies nobody under any standard.
Modules eight to ten are the content hardest to teach accurately. Module eight teaches verification without fearmongering: there is no officially investigated AI failure case in US manufacturing, so instructors quote measured shortfalls with what they were measured on and explain why accuracy misleads when failures are rare. Module nine teaches deemed exports, trade secrets, controlled unclassified information as a contract obligation and one state's biometric law, as a screen and an escalation to a named owner, never as an export determination. Module ten teaches learners to find the rule that applies to them: state safety plans, two state engineering exemptions, state and city AI employment rules, a customer requirement and a foreign rule that applies only by scope.
Modules eleven and twelve cover assessment and access. Module eleven writes scenario questions and performance tasks that test action, makes stopping the right answer where it should be, and writes rubrics a second marker applies the same way. Module twelve covers reading level and who checks a translation before it reaches a machine, noise, hearing protection and shift fatigue, accessible documents and the learner without a computer, and collecting only the information a session needs.
Module thirteen answers the questions a plant room asks. Will this take my job? Nobody has manufacturing evidence either way, and an instructor who claims otherwise is guessing. It also covers monitoring and hiring questions, saying "I do not know" and then finding out, keeping engineering, legal, export and safety calls with their owners, disclosing vendor ties and never overstating a certificate. Module fourteen prepares the Teach-Back, the capstone: one AI training segment on a fictional plant, checked claim by claim against its sources, rehearsed, delivered while observed and recorded so a correction can reach the room later. The level ships with a printable workbook and ten templates, and the examination draws forty scenario questions from a reviewed bank.
The course cites no education research it has not verified, so instructional practice is taught as practice, labelled as best practice or as this programme's recommendation. Every statement of authority is labelled: law or rule with its jurisdiction, professional standard or guidance with its issuing body and date, best practice, emerging practice, or an AI Coalition Network recommendation.
Everything here is professional education. It is not legal, engineering, quality, safety or export-control advice, and it does not replace the employer's safety-training duties, a licensed professional engineer where a state requires one, the person your quality system names, your export authority, legal counsel or the law that applies to your plant. Completing the level earns an independent educational certificate issued by AI Coalition Network with a public verification page. It is not a teaching credential, a training-provider registration or a safety-trainer qualification, it carries no professional education hours, and a session taught by a holder earns none, satisfies no OSHA, customer, registrar or scheme training requirement, and qualifies no one as a competent person, an authorised employee, an auditor or an assessor.