Research finds AI literacy alone does not predict work quality
Reviewed by Helen Jones
A KPMG and University of Texas field study reported by Harvard Business Review compared 523 early-career professionals using an AI agent on business tasks. It found that critical thinking, domain knowledge and AI literacy alone did not reliably explain who improved on the AI-only baseline; the way people worked with the agent mattered.
Managers can easily turn AI capability into a checklist of tools, prompts and courses. That may miss the practical behaviour that creates better work: framing the decision, challenging weak output and using judgment to improve the final recommendation. The study focuses on early-career professionals, so it is not a universal scorecard for every team. It is still a useful warning for leaders designing AI training: measure a real work output and the review process around it, not simply whether someone can operate an assistant.
Analysis
Choose one recurring team task and define what a good AI-assisted output must show: the source inputs, the human judgment applied and the final decision. Use that rubric in your next training exercise instead of testing prompt recall.
Source note
Pulse published by Collab365 Spaces, reviewed by Helen Jones on . Cite as "Research finds AI literacy alone does not predict work quality", Collab365 Spaces.