Background and objectives: Foundational artificial intelligence (AI) literacy is fast becoming essential to education and work, yet learners in under-resourced schools may have limited access. This feasibility study examined whether a low-infrastructure AI literacy workshop could be delivered to Grade 11 public school learners in Gauteng, South Africa, and described learners’ baseline AI-related orientations and their immediate responses and reflections after the workshop. Methods: The study adopted a pragmatic, convergent mixed-methods design. We recruited Grade 11 learners from under-resourced public schools in the Johannesburg South education district through purposive and convenience sampling. They attended a three-hour, facilitator-led workshop which did not require individuals to have digital devices. We administered anonymous paper questionnaires before the workshop (n = 237; three single-item self-report indicators) and after it (n = 220; structured evaluation items and open-ended questions). Baseline items were analysed using descriptive statistics, one-sample t-tests against the scale midpoint, and Pearson correlations; open-ended responses were analysed thematically. Results: Before the workshop, mean ratings for perceived AI clarity (M = 3.61, SD = 0.992), relevance to tertiary education (M = 3.80, SD = 0.796), and relevance to work (M = 4.17, SD = 0.795) all exceeded the neutral midpoint (p < .001). After the workshop, 211 learners (95.9%) reported greater motivation to study for their next examinations, and 218 (99.1%) reported motivation for Grade 12. Open-ended responses described aspirations, career thinking, and agency, alongside structural barriers such as limited access to devices and connectivity. Because the anonymous samples could not be matched, the baseline ratings are not meant to measure workshop effects. Conclusions: The workshop proved feasible to deliver in low-resource conditions and received positive immediate feedback. The findings support further, more rigorous testing of low-infrastructure AI literacy provision, but do not establish gains in AI knowledge, causal effects, or lasting behavioural change.