Abstract The debate on whether military humanitarian intervention and anti-atrocity norms, such as the responsibility to protect, cause perverse incentives, and provocative violence by a rebel group, has yet to reach a consensus. Specifically, existing theories are unable to fully explain why “emboldened” rebel groups provoke the government in some cases but not others. This paper reconciles this unresolved debate by arguing that it is not anti-atrocity norms but the temporary feasibility of humanitarian intervention that induces the potential rebel group to actually provoke the government. The inherent temporariness of the feasibility of large-scale intervention causes a commitment problem: An expectation of a temporary intervention shifts the balance of power toward a potential rebel and opens a quickly closing window of opportunity. A three-player game shows that, among other findings, a higher probability of humanitarian intervention alleviates the potential rebel’s incentive to fight, implying that the criticism of anti-atrocity norms is a “false charge.” The model also (i) explains why weaker groups can rebel in the shadow of external intervention and (ii) presents a novel implication that lowering the intervention cost can exacerbate the adverse incentive when a domestic commitment problem caused by the consolidation of government power is not severe.