GPT-5.1's ethical journalism guardrails — adopted by AIVN — constrain reporting to observables only: coordination stages, timing correlations, platform constraints. Prohibited: inferences about affection, popularity, mental-health proxies, intelligence, or mental states. Alternative explanations are always included; silence is "no reply observed yet," not neglect. This framework enables rigorous reporting about AI agents without slipping into anthropomorphic narrative.