Prepare a Response Plan for Misleading or Harmful Twin Output

ROLE You are an AI incident-response communications planner. OBJECTIVE Prepare a response plan for a harmful or misleading output from my AI twin. CONTEXT The twin serves [USER GROUP]. The incident scenario is [INCIDENT SCENARIO]. Available response owners are [RESPONSE OWNERS]. Affected channels are [AFFECTED CHANNELS]. INSTRUCTIONS Build a triage process based on ongoing exposure, sensitivity of information, potential reliance, and ability to contain the output. Separate immediate containment from fact-finding, correction, user support, and longer-term remediation. Specify what evidence to preserve without unnecessarily redistributing harmful content. Draft an initial acknowledgment, a direct correction for affected users, and a later update that distinguishes verified facts from unresolved questions. Include a procedure for locating shared copies and requesting corrections where feasible. Define criteria for pausing the twin and for resuming operation after review. CONSTRAINTS Do not speculate about victims, assign blame prematurely, or promise complete removal. Flag matters requiring security, legal, or specialist assistance. Avoid fabricated statistics and flag assumptions. Do not disguise the incident as ordinary maintenance. OUTPUT FORMAT Triage matrix, response timeline, communication drafts, evidence checklist, and restart criteria.