HUMAN–AI COLLABORATION IN NURSING PRACTICE, EVALUATING THE IMPACT OF INTELLIGENT CLINICAL DECISION SUPPORT, AUTOMATED DOCUMENTATION, PREDICTIVE PATIENT MONITORING, AND AI-ASSISTED CARE PLANNING ON NURSING PERFORMANCE, PROFESSIONAL AUTONOMY, CLINICAL REASONING, AND PATIENT SAFETY

The accelerated adoption of AI in nursing practice is a paradigmatic shift in the delivery of healthcare services, and intelligent systems are evolving from being a passive instrument of administration to an active cognitive agent in the 5IR healthcare system. We critically synthesize evidence across multiple domains to assess how four core AI types intelligent clinical decision support systems (CDSS), automated documentation through ambient clinical intelligence (ACI), predictive patient monitoring, and AI-assisted care planning affect nursing performance, professional autonomy, clinical reasoning, and patient safety. The results indicate that integration of the AI technology provides significant operational advantages, such as 28.8% to 30% reduction of time spent on documentation, 40% reductions in time taken for triage decisions, and 30% enhancement of efficiency in resource usage. But the positive effects of AI are balanced by profound cognitive and professional risks, including automation bias, cognitive deskilling, “never-skilling” of novice clinicians, and vigilance fatigue. The rise of the "liability sink"2 for algorithmic errors compounds asymmetrical legal risk as that algorithmic fallibility maps onto bedside nurses vicariously bearing the risk of algorithm errors absent corresponding administrative control. Moreover, patient and public safety and healthcare equity are at risk from algorithmic bias, hallucinations of large language models, and lack of adequate transparency. The GREAT PLEA system of nursing governance and participatory co-design processes provide avenues through which these risks may be addressed whilst maintaining the person-centred heart of nursing. We propose that a way forward to maximizing human–AI collaboration in nursing is by implementing compulsory “AI-free” simulation training, adopting a reasoning-first pedagogy, designing transparent explainable interfaces in partnership with frontline nurses, and developing regulatory policies that clarify liability protections. In the absence of these protections, computational intelligence could undermine rather than augment clinical judgment, moral agency, and therapeutic presence the hallmarks of professional nursing practice.