Decision Support Systems (DSS) are increasingly touted as a means to improve the efficiency and effectiveness of agricultural enterprise from variable yield mapping for nutritional inputs to autonomous irrigation scheduling. However, due to the complexity of variables and uncertainties, the statistical models that produce recommendations have limited accuracy, can be miscalibrated in situ, and/or outputs misrepresented to decision makers. In particular, AgTech manufacturers may want to hide uncertainties inherent in their technologies to drive adoption. A desire to obfuscate uncertainty makes sense because uncertainty is hard to understand, hard to calculate, hard to represent and hard to action—producing anxiety, avoidance and suboptimal decisions. Yet, users ought to be able to trust manufacturers to make reliable and useful tools with integrity. This paper presents research on the psychology of representing uncertainty and its effects on decisions that can be used to guide the design and regulation of agricultural decision support systems. Regulatory responses ought consider both the design of DSS (e.g. make explicit underlying probability distributions) and training required to operate these systems (e.g. explicit statistical training). Users must understand scenarios and consequences of decisions including low likelihood—though still possible-adverse outcomes. Manufacturers must offer product protection so that they accept responsibility and offer compensation for malfunctioning tools. Legal precedent in causality, plus research on the psychology of causality from experimental philosophy may help improve trust and adoption of agricultural decision support tools.