AI agents can be designed for ethical decision-making by incorporating ethical guidelines and constraints into their algorithms, using transparent decision-making processes, and implementing fairness, accountability, and bias mitigation techniques during their development and deployment.