Generative AI can help us write, research, plan, reason and make decisions. But as more of our thinking happens alongside AI, a harder question emerges: what happens to human agency when some of that thinking is handed over?
Our new report, Human Agency in AI Use examines this question through everyday generative AI use. One of the clearest findings is that frequent AI use alone did not signal weaker agency. The more important issue may be what people retain while using it, which is the ability to question an answer, understand the work, act independently and make the final decision.
The report develops six working dimensions of human agency that will guide the next stage of our research: independent initiation, cognitive authorship, verification and epistemic judgment, explanatory ownership, capability retention and decision authority.