How Can We Help Shape AI's Future in Healthcare?

Introduction

While understanding AI's capabilities, benefits, and risks provides a strong foundation, healthcare's digital transformation requires active participation.

The decisions shaping AI in medicine from tool procurement to daily workflows happen across all levels of healthcare. Because of this, valuable perspectives are needed from everyone:

• Clinicians evaluating tools for safe clinical integration.
• Patients navigating their care and data rights.
• Managers building effective governance and policies.
• Students preparing to navigate and shape AI-driven healthcare.

You do not need to be an AI researcher or policy expert to contribute; you just need to engage from where you are, using the influence you already have.

This unit is structured around each of the four groups this course is designed for. Find yours, take the actions that feel most relevant, and carry the wider picture with you.

Next: For Clinicians →

For Clinicians

If you work in clinical practice: as a doctor, nurse, or any other registered professional. AI tools will arrive in your department, your surgery, your ward. The question is whether you are an active participant in how that happens, or a passive recipient of decisions made elsewhere.

Develop genuine AI literacy - beyond the headlines

This course is a starting point, not a destination. AI literacy for clinicians means understanding not just what AI can do, but how to critically appraise AI tools in clinical contexts: how to read a validation study, how to identify whether a system was tested on a population like yours, how to recognise the failure modes described in Unit 3.

A growing number of resources exist for this. The MHRA's guidance on software and AI as a medical device; the NICE Evidence Standards Framework for digital health technologies; NHS England's published guidance on AI (including on ambient voice technology); and the GMC, NMC and HCPC positions on using AI in practice.

The Coalition for Health AI (CHAI) in the United States has developed practical frameworks for evaluating clinical AI. The DECIDE-AI reporting guideline provides a structured way to assess early-stage AI decision support tools. The WHO has released guidelines ensuring equitable delivery of AI in healthcare.

To support an ongoing need, our curricula and open-access resource bank are being designed specifically to bridge these educational gaps as the landscape evolves.

Ask questions when AI tools arrive in your workplace

When your institution proposes to deploy an AI tool, you have the standing to ask:

• What was this system trained on, and does that training population resemble our patients?
• What is its documented performance, and in what conditions does it fail?
• How has it been independently validated, not just by the vendor?
• What is the process for reporting concerns or errors once it is deployed?
• Who in this institution is responsible if the system causes harm?

Although they may appear obstructive, they are the questions a responsible professional should ask.

Contribute to the evidence base

Clinicians who use AI tools in practice are in a unique position to contribute to the evidence about how those tools perform in real clinical conditions. Case reports, quality improvement projects, audit data, and formal research can all emerge from clinical practice.

Consider a more formal role

For clinicians who want to engage more deeply, formal leadership routes exist in most health systems. In the UK this is typically the Clinical Chief Information Officer (CCIO) or Chief Nursing Information Officer (CNIO) role; in the US and many other countries, the equivalent is the Chief Medical Information Officer (CMIO). Fellowship programmes in clinical informatics and digital health exist in the US, UK, Australia, and elsewhere. Research positions at the intersection of medicine and AI are also expanding rapidly in academic health centres worldwide.

https://ai-podcast.nejm.org/e/from-clinician-to-chief-health-ai-officer-a-conversation-with-dr-karandeep-singh/

Talk to your patients about AI

Patients deserve to know when AI is involved in their care. As a clinician, you can help change this. A brief, honest explanation of how an AI tool is being used and that a human remains responsible for their care goes a long way toward the transparency that patients are entitled to and that trust depends on.

For allied health professionals: don't wait to be consulted

Pharmacists and allied health professionals tend to be less involved in the development of AI systems. AI tools for rehabilitation, for medication management, for diagnostic imaging, and for clinical assessment are being built at times without meaningful input from the professionals who will use them or the patients they will affect.

Engage with your professional body. If your institution is piloting an AI tool relevant to your practice, volunteer to be part of the evaluation. Your knowledge of what patients actually need from that interaction, is genuinely valuable to getting the tool right.

Next: For Patients and the Public →

For Patients and the Public: Your Rights, Your Voice, Your Choices

Discussions around healthcare AI often focus primarily on how technology impacts patients and their data. While these outcomes are vital, this perspective tells only part of the story. Far from being mere bystanders, patients and the public are essential stakeholders. They possess both the right and the valuable perspective needed to help actively shape the future of healthcare AI.

Across the EU as an example, while 91% of Member States consulted government actors and 82% engaged healthcare providers on AI governance, The broader public was consulted by 18%.

Know your data rights

Your health data is extraordinarily valuable. Both to you and to the organisations that want to use it to train AI systems. In most countries, you have legal rights over that data. In the UK the Data Protection Act 2018 and the UK General Data Protection Regulation (UK GDPR) provide robust legal protections for health data. In the United States, HIPAA governs health data privacy.

Exercising your rights begins with knowing they exist. In the UK, the National Data Opt-Out allows patients to prevent their confidential NHS data from being used for purposes beyond direct care. If you're unsure, patient advocacy organisations in your area can often help navigate data rights.

Engage with public consultations

Decisions about AI in healthcare: how it is regulated, how health data can be used, which systems receive approval are made by governments and regulatory agencies. Many of these processes include public consultation. Regulatory agencies run public comment processes on digital health and AI. Patient participation encourages a well-rounded conversation alongside established government and industry viewpoints.

Support patient advocacy in this space

A growing number of organisations are working specifically on patient rights and AI in healthcare. Disease-specific patient organisations developing AI policies for their communities, data rights advocacy groups, and research bodies that include patients as genuine partners in AI development. The two examples are just a starting point.

Talk to other patients

One of the most underrated contributions patients can make is sharing knowledge with each other. The majority of people have no idea that AI may be being used in their care, that they have data rights, or that decisions being made now will shape their future healthcare. Start a conversation.

Next: For Managers, Policy Staff, and Institutions →

For Managers, Policy Staff, and Institutions

Much of the most consequential action on AI in healthcare will happen at an institutional and policy level. The decisions made by health system managers, commissioners, regulators, and policymakers will determine the conditions under which AI tools are deployed, evaluated, and held accountable.

Make procurement decisions that the evidence supports

The commercial market for healthcare AI is large, fast-moving, and not always honest about what products can do. Responsible procurement means requiring evidence that responsible adoption depends on: independent validation conducted by parties other than the vendor; performance disaggregated by demographic; transparent failure modes; and post-deployment monitoring plans.

Build institutional governance before deployment, not after

The accountability gaps described in Unit 3 do not fill themselves. Every AI deployment should be accompanied by a governance framework that answers, in advance: who in the institution is responsible for the clinical performance of this tool? How will clinicians report concerns or errors? What is the process if the AI causes or contributes to patient harm? How will patients be informed? How will performance be monitored over time?

Waiting until a problem occurs to establish these processes is too late.

Engage in policy and regulatory processes

Health system managers and policymakers are among the most credible voices in regulatory conversations about AI. Technology companies participate actively in regulatory consultations. The perspective of those who will actually implement these systems in complex, real-world health settings is frequently missing.

Regulatory bodies globally: the MHRA in the UK, FDA in the United States, the Therapeutic Goods Administration in Australia, and others are actively developing frameworks for clinical AI. Many of these processes welcome structured input from health systems. If your institution is using AI tools in clinical care, you are already accumulating evidence that regulators need.

Take the equity question seriously as a strategic priority

The two-tier risk identified in Bonus Page of Unit 3 will not be avoided by accident. It requires deliberate institutional strategy: ensuring that AI tools deployed in your system are validated for the communities you serve, that the benefits of AI are not concentrated in already-advantaged services, and that the data contributed by underserved populations is used in ways that benefit those populations.

Next: For Students, Trainees, and Allied Health Professionals →

For Students & Trainees: Entering a Field That Is Changing

If you are at an earlier stage in your healthcare career, you are entering a field at the start of a significant transformation.

For students and trainees: build AI literacy now.

The education you are receiving was largely designed before modern AI existed. Building AI literacy now before your professional identity is fully formed is easier than doing so later.

Practically: seek out the AI in your clinical placements and ask to understand it. Pursue structured learning where it exists in your programme. Consider where your chosen specialism sits in the AI landscape: radiology and pathology are at the frontier; primary care, mental health, surgery, and rehabilitation are all in different stages of the same journey. To support this ongoing need, our upcoming curricula and open-access resource bank are being designed specifically to bridge these educational gaps as the landscape evolves.

Don't assume AI will make your skills irrelevant. The deskilling risk described in Unit 3 is real; but it is a risk of passive acceptance, not of engagement. The clinicians who will thrive are those who develop deep expertise alongside AI literacy, who can do what AI cannot (contextual reasoning, therapeutic relationships, ethical judgement), and who understand AI well enough to use it as a tool rather than a replacement for thought.

For trainees: the specific challenge of learning alongside AI

If you are in a supervised training programme, approved AI decision support tools may make it easier to reach a correct answer. But harder to develop the underlying reasoning that would let you reach that answer independently, or recognise when the AI is wrong. Use approved AI tools as a check on your reasoning, not a substitute for it. When an AI tool gives you an output, reason through what you would have concluded without it before considering a recommendation.