Guidelight AI releases its v1.0 standard to evaluate an AI system's capacity for harm rather than its tendency

The framework uses seven principles to inform deployment safeguards