In plain words
A moral agent is an entity to which moral duties can apply. Many accounts also require capacities such as understanding moral reasons, exercising sufficient control, and being appropriately held responsible for choices.
A closer look
Philosophers disagree about the exact requirements. Some emphasize intention and rational reflection; others focus on functional capacities or relationships. Adult humans with relevant capacities are standard examples. Causing an outcome is not by itself the same as being morally responsible for it.
In AI, “agent” usually has a technical meaning: a system that selects actions. That does not settle moral agency. A system can follow ethical rules or discuss right and wrong without this establishing that it has obligations or deserves blame. Whether artificial systems could qualify, and under what conditions, remains disputed.
In practice
A person who knowingly sends a misleading message may be a proper target of moral criticism. If software sends it, the moral assessment must examine the roles of designers, operators, and users rather than simply assuming the software bears the same responsibility.
A useful distinction
Moral agency and moral patienthood are different. Being responsible for actions concerns duties one has; being a moral patient concerns duties others have toward one. An entity need not possess both.