In agent-based models, agents are expected to coordinate mutual actions - to cooperate. The cooperation among agents is usually described by tools of game theory. In general, the cooperation of autonomous agents is based on information of perspective gain from cooperation. If the gain from cooperation is at least as high as the gain which agents can receive without cooperation, then this situation can be described by tools of superadditive cooperative games. The information received by agents in the case of real-world systems is not deterministic, and the use of more sophisticated tools is required. Hence, the main aim of this paper is to discuss additivity and superadditivity issues in the case of cooperative games with expectations given as Atanassov intuitionistic numbers.