AI Literacy for the Modern Workforce

Work­force AI adop­tion has out­run work­force AI judg­ment. Most train­ing on offer teach­es tool fa­mil­iar­i­ty and mea­sures com­ple­tion, and it skips the me­chan­ics of how these mod­els pro­duce their out­put that would en­able prop­er judg­ment of gen­er­a­tive AI out­puts. With­out that ground­ing, a learn­er has no basis for de­cid­ing what to hand to a model, how to spec­i­fy the work, whether to trust what comes back, or who stays ac­count­able for the re­sult. This course ex­pands from my capstone research on teach­ing AI Lit­er­a­cy through the me­chan­ics of how LLMs work. It con­sists of four mod­ules (Con­text, Ev­i­dence, Mech­a­nism, Ap­pli­ca­tion) that move a mid-ca­reer pro­fes­sion­al from the busi­ness case through the doc­u­ment­ed usage ev­i­dence into how the mod­els ac­tu­al­ly gen­er­ate text, and fi­nal­ly into ap­plied prac­tice. The plat­form was or­ches­trat­ed to be cus­tomized so the in­ter­ac­tion de­sign and data cap­ture would not in­her­it the lim­i­ta­tions of cur­rent au­thor­ing tools. The pro­gram and build took rough­ly 210 hours of solo work. The needs analy­sis, the eval­u­a­tion frame­work, and the build records are pub­lished below.