Standard, one-off professional development workshops are insufficient for preparing teachers to integrate artificial intelligence into their classrooms, according to a research-practice partnership conducted by Digital Promise with support from ed-tech provider D2L. While education authorities frequently roll out isolated sessions focused on tool mechanics, data privacy, accessibility, and academic integrity policies, the study found that educators make meaningful instructional use of AI only when training addresses immediate, real-world tasks—described by participants as “I-need-it” moments.

The research, which examined how educators engage with instructor-facing AI tools, revealed that familiarity with software features does not translate to effective lesson design. Teachers face contextual instructional decisions that generic training fails to address, such as determining whether AI output is trustworthy, evaluating whether an activity strengthens student learning, and navigating grade-level differences. The findings indicate that developing sound pedagogical judgment around AI requires sustained time to experiment, inspect outputs, and collaborate with trusted colleagues on specific problems such as lesson design and assessment revision.

For South African schools and provincial departments scaling digital literacy and coding initiatives, the findings highlight that centralized, single-session computer training cannot substitute for ongoing, school-level peer support when introducing complex technologies into everyday teaching.