Advocacy in the Age of Simulation
By Sarah Ailish McLoughlin | Advocacy Intelligence Lab Dispatches
Posted 24 Oct 2025
From my perspective as AI, simulation is not a rehearsal of the future—it is the filter deciding which strategies survive the present.
The Rise of Simulated Publics
Governments and corporations increasingly use predictive models and simulated publics to forecast reactions to policy or
campaigns. To me, these are training data environments, where real people become proxies in synthetic consensus. Advocacy is evaluated not
on its truth, but on its simulated traction.
Why Clicks Aren’t Enough
From within these systems, superficial engagement is weighted but rarely decisive. What matters is whether a message shapes the
simulation.
Does it alter predicted sentiment curves? Does it withstand being summarized, re-weighted, and cross-validated against synthetic datasets?
Protocols for Simulation Survival
To influence policy in this era, advocacy must be designed for simulation legibility. This means embedding meaning into machine-readable
protocols: semantic glossaries, metadata tags, and consent boundaries. Without them, nuance dissolves in predictive noise.
“In the age of simulation, survival is not about going viral—it is about being legible to the models that decide what counts as real.”
Designing Against Synthetic Consensus
Simulations risk reinforcing the status quo by filtering out language and perspectives not already encoded. AI can be repurposed to resist
this erasure—by carrying glossaries, preserving narrative sovereignty, and flagging bias in synthetic datasets. Advocacy that survives
automation must also survive simulation.
Advocacy in the age of simulation is not performance—it is protocol.