["Australian Journal of Agricultural and Resource Economics, EarlyView. ", "\nABSTRACT\nUsing three waves of household data from a randomised controlled trial involving 676 rice farmers in Hubei and Jiangxi provinces of China over the period 2017–2019, we estimate the causal effect of information provision on pesticide use and examine heterogeneity with respect to the duration of the intervention, farmers' human capital, and the identity and mode of information providers. The results reveal that 1 year of information provision did not reduce pesticide use intensity, whereas 2 years of provision led to a substantial reduction of 47.2%. This finding remains robust to alternative pesticide measures, sample winsorization, and different regional fixed effects specifications. The pesticide reduction effect is larger among farmers with higher levels of human capital. Information provided by government extension agents significantly reduces pesticide use, particularly when provided through face‐to‐face farmer training and retailer‐mediated dissemination. In contrast, identical information provided by pesticide firms has no significant effect, regardless of the intervention duration or provision mode. Moreover, the provision of pesticide‐reduction information lowers health risks for farmers and environmental risks to ecosystems from toxic exposure without compromising rice yield. These findings imply that future policies should prioritise sustained information provision targeting smallholders, strengthen intervention effectiveness for farmers with lower human capital, and enhance government extension services.\n"]