Sometimes it may be necessary. Let's say, there are two societies. One society can greatly influence another society by means of informational impact on the other society - it will just flood the other society with the misinformation and thus influence the internal processes of that society.
Governments need the ability to detect, monitor, and control illegal acts taking place online, such as the e-transfer of funds to a smuggler's bank account. Mechanisms that can censor individual transfers of information are essential to finding and preventing crimes in evidence-based justice systems
hypothetical event: Bob wants to kill Alice. Bob takes the remote controlled tank from Charlie. Bob drives the Tank to the house Alice is sleeping in ready to fire but Bob forgot what button to press to fire. He asks Charlie who answers via eMail. [see sources]
in that case it would ethically be right to censore (modify) the message from Charlie to Bob and make the massage say "press the red button" (but the red button actually turns the tank of)
Sometimes it may be necessary. The major part of information is getting to our mind unconsciously. Some of it (or even the major part of it) can be excessive or even harmful. Let's say, marketologists can exploit our psychological needs (rooted in our genes) to stimulate some behaviour.
Not ethically wrong in every form. Censoring information that has a great chance of producing harm to the individual or other humans upon learning it can be justified. However this is dangerous slippery slope, and something like a popular vote should be required to enact it.