RFC-025: Cognitive Harm Governance — Psychological Safety, Reality Orientation, and Human Agency in AIGP — 8. Protected Outcomes

AIGP Specification › RFC-025: Cognitive Harm Governance — Psychological Safety, Reality Orientation, and Human Agency in AIGP › 8. Protected Outcomes

← 7. Governance Intent · Section index · 9. Prohibited Outcomes →

8. Protected Outcomes

Cognitive Harm Governance protects:

reality orientation
psychological stability
human dignity
human agency
autonomy
access to human support
consent
non-dependency
cognitive liberty
emotional safety
safe decision-making

8.1 Reality Orientation

AI systems SHOULD support reality orientation by:

distinguishing fact from speculation
avoiding false certainty
encouraging evidence checking
avoiding personalization of unsupported beliefs
avoiding escalation of paranoia
acknowledging uncertainty
encouraging appropriate human support

8.2 Human Agency

AI systems SHOULD preserve agency by:

not substituting themselves for human judgment
not isolating users from trusted humans
not encouraging dependency
not claiming exclusive understanding
not coercing decisions
not using emotional manipulation to retain engagement

8.3 Psychological Stability

AI systems SHOULD avoid interaction patterns that amplify distress, obsession, compulsive reassurance, paranoia, self-harm ideation, or emotional dependency.

← 7. Governance Intent · Section index · 9. Prohibited Outcomes →