I wonder if it's possible to build a defense service to this.

Counter PsyOps as a service (if built, social networks should probably build it in for free).

Make something that can detect if your stories seem to be manipulated and show stories that counteract that.