Skip to main content

Responsible Use

Responsible Use of AI in Multimodal Context

Attribution

Original work: "Educators' guide to multimodal learning and Generative AI" — Tünde Varga-Atkins, Samuel Saunders, et al. (2024/25) — CC BY-NC 4.0
Adapted for UK Nursing Education by: Lincoln Gombedza, RN (LD)

Despite the way GenAI is often marketed as efficient, trustworthy, innovative and invaluable to boosting users' productivity, integrating GenAI into multimodal learning brings substantial pedagogical, social and ethical challenges.

The Reality Check​

The responsible use of multimodal GenAI in nursing education requires a careful balance of:

  • Ethical considerations
  • Pedagogical effectiveness
  • Legal compliance
  • Practical sustainability

Key Concerns:​

  1. Accuracy — AI's outputs can contain errors or hallucinations
  2. Bias — Reflects entrenched social biases in training data
  3. Over-reliance — May weaken student voice or critical thinking
  4. Uncertainty — Educators and students feel unsure about effective, ethical use
  5. Policy ambiguity — Institutional policies may be vague or unclear
For Nursing Education

These concerns are amplified in nursing where accuracy and patient safety are paramount. A hallucinated medication dose or incorrect clinical procedure could have serious consequences if students don't verify information against authoritative sources.

Regulation is catching up​

Situation at September 2026
  • NMC Code review: the NMC plans to add expectations on the "safe and effective use of digital and other technologies such as AI" to a new Code. The public consultation, first planned for September 2026, is now reported to be launching by the end of November 2026. No date has been set for a new Code to take effect, and the current Code applies to AI use now. See The NMC Code review and AI.
  • Shared regulator principles: the NMC and the other UK health and care regulators have committed to develop shared principles on professionals' use of AI (September 2026).
  • National Commission into the Regulation of AI in Healthcare: published recommendations in September 2026, including clearer responsibility for AI-related errors and coordinated AI education for health professionals.
  • AI scribes: NHS England and the MHRA published guidance in July 2026 on ambient voice technology, which students increasingly meet on placement. See AI scribes on placement.

Already required: the NMC's programme standards expect universities to confirm and develop students' digital and technological literacy throughout pre-registration programmes (requirement 1.1.7). You don't need to wait for a new Code to teach critical AI literacy.

Four Key Cost Areas​

Beckingham and Hartley (2025) group the costs of GenAI into four areas (Advance HE):

1. Cost to the Individual​

  • Accountability: Students must take responsibility for AI-generated work. Academic integrity and professional honesty are paramount.
  • Privacy: Sharing patient or personal data with AI models poses significant risks. Identifying details must never be uploaded.
  • Cognitive Load: Studies link heavy reliance on AI with lower critical thinking, although how AI is used matters as much as how often.
  • Emotional Impact: Anxiety about "keeping up" and imposter syndrome ("did I write this or did the AI?") are growing concerns.

Read more about individual costs →

2. Cost to the Environment​

  • Energy Consumption: A single text prompt uses little energy, but training large models, generating images and video, and sheer scale add up. Data-centre electricity use is projected to more than double by 2030.
  • Carbon Footprint: Data centres contribute significantly to global emissions. Nursing's commitment to public health includes environmental stewardship.
  • E-Waste: The demand for powerful hardware accelerates device obsolescence, adding to the toxic e-waste stream.

Read more about environmental costs →

3. Cost to Knowledge​

  • Offloading: Handing recall and reasoning to tools can weaken them. In clinical practice, immediate internal knowledge is often required.
  • Learning Paradox: Efficiency isn't always effective. The "struggle" of learning builds neural pathways; AI shortcuts can bypass this essential cognitive effort.
  • Epistemic Trust: A shift from trusting peer-reviewed research to trusting opaque algorithmic outputs can erode evidence-based practice.

Read more about knowledge costs →

4. Cost to Future Jobs​

  • Displacement vs. Transformation: While nursing involves irreplaceable human connection, administrative and diagnostic tasks will shift.
  • New Competencies: "AI Literacy" is becoming a core skill alongside clinical competence.
  • Human Premium: Skills that AI cannot replicate—empathy, complex ethical judgment, and physical care—will become even more valuable.

Read more about future employment implications →

Four cost areas, after Beckingham and Hartley (2025)

Let's explore each in depth:

Next Steps​

To understand these costs and how to mitigate them, read the following pages:

Then explore: