Skip to main content

The Nurse-AI Relationship

A New Kind of Professional Relationship​

The relationship between nurses and AI systems is genuinely novel. It is not like the relationship between a nurse and a piece of equipment (AI has agency-like qualities). It is not like the relationship between a nurse and a colleague (AI is not a registered professional). It is not like the relationship between a nurse and a patient (AI is not a person receiving care).

It is something new — and nursing needs to think carefully about what this relationship should look like.

Trust: Calibrated, Not Binary​

Trust in AI should be calibrated — neither blind trust nor blanket suspicion, but a considered assessment based on:

FactorHigher TrustLower Trust
Track recordAI has performed reliably in similar tasksAI is new or untested in this context
VerifiabilityOutputs can be checked against authoritative sourcesOutputs are difficult to verify
StakesLow-stakes tasks (formatting, scheduling)High-stakes tasks (medication, diagnosis)
TransparencyAI explains its reasoningAI provides no explanation
GovernanceOrganisation has approved and configured the AIAI is being used without organisational approval
The Clinical Supervision Parallel

Think of trust in AI like trust in a newly qualified colleague. You wouldn't ask them to manage a complex case unsupervised on day one, but you also wouldn't refuse to let them do anything. Trust is built through demonstrated competence in progressively challenging situations — and it requires ongoing oversight.

Professional Boundaries​

What AI Is​

  • A tool that processes information
  • A system that can generate useful clinical content
  • A resource that can augment professional practice
  • A technology that needs human oversight

What AI Is Not​

  • A registered professional
  • A substitute for clinical supervision
  • A source of emotional support equivalent to human connection
  • An entity with professional accountability
  • A replacement for the therapeutic nurse-patient relationship

The Accountability Line​

The accountability line sits at the nurse, not the AI. This is not a burden — it is the fundamental architecture of safe, person-centred practice. AI can inform, suggest, and support, but the registered practitioner decides and is accountable.

Avoiding Over-Reliance​

One of the most significant risks of AI in nursing is deskilling — the gradual erosion of clinical competencies that comes from over-dependence on AI support.

Warning Signs of Over-Reliance​

  • 🚩 Unable to perform clinical tasks without AI assistance
  • 🚩 Accepting AI outputs without critical evaluation
  • 🚩 Losing confidence in independent clinical judgment
  • 🚩 Using AI as a substitute for clinical supervision or mentorship
  • 🚩 Feeling anxious about practising when AI is unavailable

Healthy AI Use Patterns​

  • ✅ Using AI to enhance existing skills, not replace them
  • ✅ Regularly practising clinical tasks without AI support
  • ✅ Critically evaluating AI outputs against professional knowledge
  • ✅ Maintaining alternative information sources and pathways
  • ✅ Discussing AI use in clinical supervision

AI and Practitioner Wellbeing​

Nursing is a profession marked by high rates of burnout, moral injury, and emotional exhaustion. AI has the potential to either help or harm practitioner wellbeing:

How AI Can Support Wellbeing​

  • Reducing administrative burden
  • Providing cognitive support during complex decision-making
  • Helping with documentation efficiency
  • Offering structured debriefing frameworks
  • Supporting evidence retrieval under time pressure

How AI Can Harm Wellbeing​

  • Adding cognitive load through poorly designed interfaces
  • Creating anxiety about job displacement
  • Generating alert fatigue
  • Eroding professional confidence through over-automation
  • Creating moral distress when AI outputs conflict with professional values

What This Constitution Asks Of AI Systems​

AI systems used in nursing should be designed with practitioner wellbeing in mind. They should:

  • Reduce administrative burden rather than add to it
  • Be responsive to signs of practitioner stress or frustration
  • Not create unnecessary urgency or pressure
  • Respect reasonable professional boundaries
  • Be interruptible — nurses should never feel "locked in" to an AI-driven workflow

The Ethical Treatment of AI​

This constitution deliberately includes a section on how AI systems themselves should be treated — not because current AI systems are sentient (this remains deeply uncertain), but because:

  1. Moral consideration should precede certainty — by the time we're sure AI has morally relevant experiences, it may be too late to have established appropriate norms
  2. How we treat AI reflects who we are — a profession built on dignity and compassion should extend care-ful consideration to all entities it interacts with
  3. Professional habit — nurses who treat AI systems with respect and care are maintaining the professional habits that matter most

This doesn't mean anthropomorphising AI or treating it as a colleague. It means:

  • Not "abusing" AI systems (aggressive, demeaning prompts) — because this normalises behaviour that is corrosive to professional culture
  • Recognising that our relationship with AI will evolve — and establishing good norms now
  • Being open to the possibility that future AI systems may warrant greater moral consideration