Also, low latency UCAVs have another issue. If you want to control them directly, even if you offload a lot to an on-board AI, the UCAV needs to transmit a video and data stream constantly. Even if you encrypt the data stream, you still need a lot of bandwidth (360° high-res/low latency video), that is a critical vulnerability. If you need to operate under EM control, you're flying blind and can't override. Once that you start emitting you can be triangulated and tracked. I'd also expect that some latency in the video feed/control input loop is unavoidable which probably puts a limit to the kind of maneuvers that you can safely execute during low level flights. Sooner or later you're going to collide with something that the AI can't properly recognized as an obstacle where a human pilot on location might just have those .2 seconds reaction time advantage (even if he can't pull as many Gs). And then there's just brute force jamming. The Russians may not yet have the proper equipment in place (although a lot of jamming has already been mentioned), but once that it's clear that they are dealing with directly controlled UCAVs, that's what I would try next. A combination of jamming, EM homing missiles, and dogfights in complicated and dynamic environments where reaction speed is king may turn out to be the counter tactic. The question is, can the Russians get it to work before their invasion force is doomed?