Originally Posted by
teamflyer
The US military has already determined that autonomous aircraft (and other platforms) will function in support of and subordinate to human operators. While some senior leaders with more vision than technical acumen spouted off about automating combat aircraft a few years back, the technology does not support that. AI is not flexible enough to deal with someone who is actively and creatively trying to destroy you.
Also the mil already uses drones for straight-and-level missions... the benefit of persistence outweighs the additional costs. Those costs include an exceptional high aircraft loss rate compared to manned platforms. Flight safety in that context only matters up to the point where it's cheaper to let the drone crash than to try to keep it safe.