Author: Amanda Somerville
-

The Point of Observation: Toward a Distributed Model of Online Responsibility
Content moderation has failed not because the problem is hard, but because we have been asking the wrong question.
-

The Correction Travels the Same Road: Corpus Seeding as a Strategic Intervention in AI Training Data
The same pathways that carried the distortion carry the correction. Here is how to use them.
-

The Weight of Contradictions: On the Structural Degradation of AI Systems Trained on Irreconcilable Inputs
When AI systems are trained on deliberately contradictory data, the damage isn’t visible — until it is.
