Navigating AI in Health Education

Original Editor - Michael Rowe Top Contributors - Jess Bell

Introduction

"As educators, we're experiencing something unprecedented: watching our professional landscape transform in real-time while trying to maintain the standards and values that define quality health education."

The questions surrounding AI use reflect deeper anxieties about maintaining academic integrity, preparing competent healthcare professionals, and adapting to technological change that feels simultaneously exciting and overwhelming.

These aren't just policy questions—they're fundamentally about how we understand teaching, learning, and professional development in an AI-enhanced world.

Data and Privacy

The Ethical Landscape

The question of data protection in AI systems touches something fundamental about the role of healthcare educators. Educators and teachers act as temporary custodians of deeply personal information about their students' learning journeys, struggles, and growth. This responsibility has always existed, but AI amplifies both the opportunities and the risks in ways that demand thoughtful consideration.

Key observations:

• privacy violations in educational AI often result from convenience rather than malice
• students trust educators with vulnerable information that extends far beyond academic performance
• the consequences of data misuse compound over time in ways we don't immediately recognise
• clear guidelines protect both students and educators from unintended harm

What data shouldn't be shared with AI?

A useful rule of thumb when deciding what data to share with AI might be to ask if you would be comfortable including your grandmother's personal information in an AI training dataset accessible to unknown future users.

Because AI systems depend on extensive data, educational institutions are at risk of data breaches and the misuse of students' personal information. Therefore, certain data should never be shared with public AI platforms, such as:

• student identification information (including names, student IDs, contact details)
• academic records, grades, and performance assessments
• personal health information, including disabilities and accommodations
• financial information and family circumstances
• behavioural observations that could stigmatise or label students
• any content that could identify specific individuals when combined with other information

These restrictions aren't just about compliance—they are about maintaining the trust that enables authentic learning relationships. When students know their personal information is protected, they are more likely to engage honestly with difficult topics, admit confusion, and take intellectual risks essential for professional development.

Consequences that extend beyond policy violations

The immediate consequences of data protection violations are well-documented and include: legal liability, policy infractions, professional sanctions, and institutional reputation damage. But the deeper consequences may prove more significant:

• students may become more guarded in their learning when they sense privacy uncertainty
• faculty may avoid beneficial AI collaboration due to unclear data guidelines
• educational innovation may stagnate when risk aversion dominates decision-making
• professional relationships may suffer when trust around data handling erodes

Practical guidelines for safe AI collaboration

Rather than approaching data protection as a series of prohibitions, it can be helpful to develop positive frameworks for safe AI use that enable rather than constrain educational innovation.

Safe data practices include using aggregated, anonymised information for curriculum development and creating fictional scenarios that reflect authentic diversity without real identifiers. It is also helpful to generate teaching materials using general professional knowledge rather than specific cases and to collaborate with AI using your expertise rather than student-specific information.

Boundary-setting strategies include developing institutional templates for AI-appropriate content creation and creating quick, contextually relevant decision-making rubrics for data sharing questions. It is important to establish regular review processes for AI tool usage and to build faculty confidence through guided practice with safe scenarios.

Building trust

The most effective data protection approaches don't rely solely on rules and restrictions. Instead, they develop professional judgement and create cultures where privacy protection becomes second nature.

Key observations:

• faculty need practical experience with AI tools to understand data implications
• students benefit from knowing that privacy is prioritised and understanding how their information is protected
• transparent processes build confidence in AI-enhanced education
• regular dialogue prevents policy drift and maintains vigilance

Just as healthcare professionals develop intuitive patient confidentiality practices, educators can develop instinctive data protection habits. This requires moving beyond checklist compliance toward a deep understanding of why these protections matter and how they serve our educational mission.

Emerging considerations in educational AI

Traditional data protection frameworks, developed for static information systems, sometimes struggle to address the dynamic, generative nature of modern AI tools. We need evolving approaches that address contemporary challenges while maintaining core protective principles.

Contemporary challenges requiring attention:

• AI systems that learn from user interactions and potentially retain information
• cross-platform data sharing when using multiple AI tools
• international data transfer when using global AI services
• long-term data retention in AI training datasets
• aggregated data that could become identifying when combined with other sources

Innovative institutions might consider developing adaptive frameworks that can respond to technological change while maintaining protective principles. This requires building institutional capacity for ongoing evaluation rather than relying on static policies.

Data and Privacy Reflections

Working with AI in healthcare professions education shows that data protection isn't only about avoiding negative outcomes—it's about creating conditions where authentic learning can flourish. When students know their information is protected, when faculty understand clear boundaries, and when institutions support innovative practice within ethical frameworks, educational quality improves significantly.

Several factors can give us confidence going forwards:

• healthcare educators' existing commitment to confidentiality provides a strong ethical foundation
• clear guidelines enable rather than constrain educational innovation
• students appreciate transparency about how their information is protected
• professional communities can develop sophisticated approaches to emerging challenges

The goal isn't perfect risk elimination—it's developing professional competence in managing evolving technological capabilities while maintaining the trust relationships essential for effective education. Students deserve both innovation and protection, and thoughtful approaches to AI data handling can deliver both.

This balance requires ongoing attention, professional dialogue, and adaptive frameworks that can evolve with technological change while maintaining core ethical commitments. Educators have the opportunity to model the kind of thoughtful, principled technology integration we hope to see throughout healthcare practice.

Copyright

Teacher’s Perspective: Balancing Innovation and Responsibility

Educators can legitimately use AI for case study development, lesson planning, assessment creation, and curriculum design—provided they verify the accuracy of content and document the process transparently.

However, this focus on technical compliance misses something important. Namely, the benefits that might be gained by asking different questions, such as:

• how does AI assistance enhance and support rather than replace pedagogical expertise?
• what verification processes protect patient safety while enabling innovation?
• how do we model the collaborative, evidence-based practice we want students to develop?

AI tends to work best when it amplifies existing expertise rather than substituting for it. Instead of focusing time and energy on what might be replaced, it's important to look for what can be updated and enhanced. The most successful AI-generated content emerges from clear educational objectives and thorough clinical knowledge.

When can teachers claim authorship of AI-assisted work?

An educator can’t “pretend ownership” of content written by AI. Current standards suggest that authorship requires substantial intellectual contribution beyond basic prompting. This means providing conceptual frameworks, pedagogical design, clinical expertise, and significant modification of AI outputs.

Key observations:

• documentation of thinking processes matters more than arbitrary contribution percentages
• professional judgement about meaningful contribution develops through experience, not rules
• healthcare practice is inherently collaborative and maybe our teaching should model this

Rather than defensive attribution, we might embrace transparent collaboration that demonstrates sophisticated AI partnership skills; skills that our students will need throughout their careers, as we see increasing use of AI in clinical practice.

Is disclosure necessary when AI is used?

Disclosure is necessary when AI is used. But disclosure serves learning objectives beyond mere compliance. When transparent AI use is modelled, important evaluation skills that students need are demonstrated.

Effective disclosure practices can include specifying which tools were used and for what purposes, documenting verification and modification processes, framing disclosure as a demonstration of professional competency and using transparency as a teaching opportunity rather than a defensive measure.

The goal of disclosure isn’t to protect ourselves from accusations. It is to prepare healthcare professionals to collaborate effectively with AI while maintaining clinical reasoning independence.

What are the consequences of non-compliance?

The immediate risks of non-compliance include academic misconduct charges, copyright concerns, and policy violations. More seriously, undisclosed AI use in health education could compromise patient safety if clinical information proves inaccurate.

However, we should also consider the deeper consequences of excessive caution:

• students may learn to hide their AI use (because the benefits of use are likely to be increasingly significant, so they will use it) rather than develop skilled collaboration
• faculty may miss opportunities to model professional technology integration
• graduates may enter AI-enhanced healthcare environments unprepared for effective practice

Attempts to avoid risk can sometimes create greater long-term hazards for both students and patients.

Student's Perspective: Developing Professional Competency

When is student AI use permissible versus prohibited?

Most institutions currently permit AI for research assistance, concept clarification, and draft improvement (although the specific details will vary by institution), but prohibit substantial content generation, especially for assessed work. This binary approach, while administratively convenient, doesn't reflect healthcare practice reality.

Students actually need to learn how to evaluate AI-generated information for accuracy and bias and to see when AI collaboration enhances versus undermines professional decision-making. They also need to understand how to maintain critical thinking while leveraging AI capabilities and gain insight into the ethical frameworks for AI use in education and patient care contexts.

Perhaps the quality of an “AI collaboration” should be assessed rather than eliminating AI entirely from academic work.

What AI use is acceptable in academic writing and research?

In general, AI use is acceptable to assist research, help with structural organisation, enhance clarity and explain concepts.

Typically, AI use is unacceptable when it replaces critical analysis, provides ready-made arguments and substitutes for original synthesis.

General observations:

• the most effective AI collaboration involves iterative questioning and critical evaluation, mirroring how students work closely with research supervisors
• students develop stronger analytical skills when they evaluate and improve AI outputs
• assessment design matters more than AI policies—good assignments make cheating irrelevant (admittedly, this is a difficult problem)
• professional writing skills improve when AI handles the mechanics of writing, while students focus on reasoning

Broader implications for professional development

As AI technology becomes more prevalent in many sectors, graduates will need to understand AI principles to succeed in their chosen fields. AI education policies "can provide students with the necessary knowledge and skills to work with AI in a professional capacity."

Student AI literacy will soon directly have an impact on patient care. Therefore, poor AI collaboration skills in education are likely to become patient safety risks in practice.

Essential competencies for healthcare professionals include:

• digital literacy as a core professional skill
• ethical reasoning for AI use in collaboration with patients
• lifelong learning strategies that incorporate AI collaboration
• critical evaluation of AI-generated information

"We're not just teaching current students—we're shaping the healthcare professionals who will serve our communities in the future."

Moving Beyond Control Paradigms

It seems that institutions that focus on control and prohibition are more likely to force AI use by students underground, while those emphasising competency development are likely to create more skilled, ethical practitioners.

Effective approaches share common characteristics:

• clear guidance for acceptable use-cases rather than rigid prohibition 
• assessment innovation that evaluates the quality of students’ collaboration with AI
• faculty development in AI competency alongside students
• community building for sharing effective practices

Moving toward sustainable integration

The questions we're addressing reflect a transitional moment in healthcare education. We can approach this transition defensively, trying to maintain existing practices, or we can embrace it as an opportunity to enhance our educational effectiveness while preparing students for AI-integrated healthcare practice.

Key insights from early adopters:

• transparency and skilled collaboration may matter more than restrictions
• professional competency includes AI literacy in education and healthcare contexts
• educational quality improves when AI amplifies rather than replaces human expertise
• patient safety requires healthcare professionals who can collaborate effectively with AI

Embracing Uncertainty

"Working with AI in health education has challenged many of my own assumptions about teaching, learning, and professional development. Initially, I also approached AI with caution, focusing on risks and limitations. Over time, I've discovered that thoughtful AI integration can enhance educational quality while developing essential professional competencies."

It is important to recognise that students will use AI regardless of policies. It's simply too useful for them to ignore. Healthcare educators must choose between guided, skilled use that serves patient care, or hidden application that bypasses professional development opportunities.

Several factors can give us confidence going forwards:

• healthcare educators' commitment to patient safety provides ethical grounding for AI integration
• students demonstrate a remarkable capacity for sophisticated AI collaboration when properly guided
• professional communities can adapt successfully to technological change while maintaining core values
• educational innovation often emerges from periods of uncertainty and challenge

Questions about AI usage, while reflecting legitimate concerns, represent an opportunity to strengthen rather than constrain healthcare professions education. By embracing thoughtful AI integration, we can prepare healthcare professionals who are more skilled, more adaptable, and ultimately more effective in serving their communities.

The future of healthcare depends not on avoiding AI, but on learning to collaborate with it skillfully and ethically. Educators have the opportunity—and responsibility—to lead this transformation.