Thanks for that. The goal is to design a system that prevents abuses. A technological solution to the ape problem would be helpful.
As an excellent example, Certificate Transparency has almost completely mitigated the potential abuse of compromising a certificate authority and using it to MITM traffic. Similarly, "binary transparency" or "software transparency" will hopefully eliminate the abuse of delivering a "special binary" to just one person that others have not received.
Part of the threat model is the belief that any system with a backdoor has any hope of "preventing abuses". The backdoor is the abuse, leaving aside all the misuses of it that will happen.