Clinical Reach into the Cognitive Space (CRITiCS): outline conceptual framework for safe use of generative artificial intelligence in mental health decision-making.

Advances in generative artificial intelligence, particularly through large language models, like GPT-4, have opened opportunities to develop cognitive agents to enhance clinical productivity, especially in complex secondary and tertiary care settings. However, as artificial intelligence begins to occupy the cognitive space traditionally held by human clinical reasoning, transparency becomes a significant concern. Unlike human decision-making, artificial intelligence-generated outputs may not be traceable to a transparent chain of clinical reasoning, potentially impacting safety if used without adequate 'clinician reach' into the reasoning space of artificial intelligence. We highlight the need for a consensus framework to guide the responsible use of generative artificial intelligence in mental healthcare, which, it is argued, has cognitive demands and features distinct from physical medicine. We propose such a framework, Clinical Reach into the Cognitive Space (CRITiCS), to support clinician involvement in the deployment of these technologies. This paper aims to spark dialogue and interest in both the clinical and artificial intelligence development communities.