This standard provides a framework for formulating and evaluating how well the behavior of autonomous driving agents aligns with human intentions. This framework specifies evaluation setups and related data, methods, and metrics. The standard defines a minimum set of scenarios, in which human intentions are encoded through agent behaviors. In each scenario, the standard defines a set of behaviors, where different behaviors reflect different levels of alignment between human expectations and agent behaviors. For instance, in a scenario, we might have three behaviors, A, B, and C, where A is the most expected behavior aligned with human intentions, B is not an expected behavior but acceptable, and C is not an acceptable behavior. Based on that, the minimum requirements for alignments and system interpretability are provided from an evaluation process perspective. This standard applies to autonomous agents on the road, which are implemented by Artificial Intelligence (AI) algorithms, machine learning, or other applicable methods. Complying with this standard helps to achieve that the fundamental performance of self-driving vehicles aligns with human intentions.