My point is - by employing domain fronting against censorship, you bet that the adversary will not ban the service you're using as a front. But they very well might just do that. At this point, everyone else using the service suffers. So that service, by refusing to be used as a domain front, is not just protecting its own interest - it's protecting interests of all the others who depend on it. You, on the other hand, are unilaterally putting those other people at risk. This does not make you a hero, it makes you a villain (even if a lesser one).
The domain fronting could be set up in a way that doesn't spoof domains, and the risks would be exactly the same. The spoofing is a red herring. The issue is the mere idea that a censor would be unable to tell what domain a connection is for. The actual thing that puts people at risk is ridiculous to attack on a moral basis. It's the same as just existing in a crowd. Not grabbing someone to be your shield.
It's like helping Assad to find targets where people with opposing views live.
But if they were going to nuke the city? Fuck them, don't negotiate, it is absolutely not the fault of any group that is merely located somewhere inside the city.