It's more likely that you aren't seeing a video feed from its "eyes." Instead you're seeing multiple things: the video feed and the logs of its internal thought processes. The movie chose to visually display both things on screen for you to see simultaneously.
Your idea that it may be taking orders would make sense for humans to receive and react to input (like in Agents of S.H.I.E.L.D. with the mechanical eyes) but this wouldn't make sincesense for a machine, because it doesn't need to read from video to get input when it could simply receive and process that input on its own. That would involve receiving data, overlaying that on the video feed, and then reading that again, which is entirely redundant and wasteful.
Since it's not human, there is no such thing as a distinct difference between the data of a visual feed or of a log of its thought processes. It's all just a bunch of 1
's and 0
's inside the processor, so the movie had to display something visually, and it chose to show the video feed and the thought process logs which were both things that can be shown visually.