AI Ethics in Healthcare: Doing it Right

Wim Van Hecke - founder/CEO icometrix Artificial intelligence in healthcare is often portrayed as the counterpart of the human touch. However, AI is not just about algorithms and automation, it’s about people. It’s about a neurologist making a tough call on a patient’s diagnosis or treatment. It’s about a radiologist looking for subtle changes in a complex MRI scan. It’s about a person who’s living with a neurological disorder and deserves the best possible care. Beyond the neural networks, the 0’s and 1’s, advanced pattern recognition, cloud computing, etc., AI supports people. AI allows any clinician, located anywhere in the world, to make decisions for individual patients with a similar accuracy and quality as global key opinion leaders and specialized experts. Hence, AI has the power to democratize care and impact equity and equality of care. 

With the enormous impact AI can make on people’s lives comes an immense responsibility. At icometrix, we don’t take this responsibility lightly. As AI transforms radiology and neurology at the speed of light, we make sure it does so in a way that clinicians can confidently make these important clinical decisions. Below, some of our guiding principles are listed:

Explainable: Clarity builds trust

Trust in AI starts with understanding. That’s why we ensure that every icobrain solution is clear about what it does, how it does it, and what its limitations are. AI isn’t magic - it’s math. Our AI-generated reports provide the necessary context, clearly showing what the system has found and why. icobrain is built on the principles of explainable AI.  It’s not a black box, but a transparent tool that healthcare professionals can confidently integrate into their clinical workflow.

Inclusive: AI that works for every brain

AI algorithms that only work for some patients aren't good enough. From the very beginning, we made a conscious choice: to train icobrain not on polished clinical trial data, but on real-world data, with all its imperfections. That means embracing variability in scanner manufacturers, patient demographics, and a broad spectrum of neurological conditions. It's messier, but it’s what real clinical care looks like.

Clinicians need to trust that an AI tool won’t fail when a case is atypical. By training on data that captures that complexity, we build AI that generalizes better and adapts to the unpredictable nature of clinical practice.

Equitable: AI that works for every brain, everywhere

Access to high-quality care, diagnostics, and treatment is a fundamental human right as stated in Article 25 of the United Nations Universal Declaration of Human Rights. Yet in practice, that right is far from guaranteed.

Care for brain disorders is becoming increasingly specialized, especially with the arrival of new disease-modifying treatments for Alzheimer’s disease, which require early diagnosis and ongoing safety monitoring. But not everyone has access to that level of specialized expertise.

That’s where we are trying to make a difference. We enable non-specialists to interpret brain MRIs for Alzheimer’s disease with accuracy and confidence, matching the performance of top global experts. And with the CPT-III code for icobrain now in place, nearly all centers across the U.S. can access icobrain for their patients.

In doing so, we're helping close the gap in Alzheimer’s care, which is also a key focus of our recently announced collaboration with the American College of Radiology. 

Personalized: AI that is tailored to everyone’s needs 

Every brain is different, and so is every brain disorder. To truly support clinical decision-making and improve outcomes, AI must deliver insights that are both precise and relevant.

That’s why each icobrain report is developed specifically for a particular neurological condition, focusing only on what matters clinically and measuring it with the highest possible accuracy. For example, brain structure volumes are compared not just to a general population, but to large, age- and sex-matched groups of both healthy individuals and people with the same condition. This dual comparison makes the results more meaningful and easier to interpret for individual patients (van Nederpelt et al., 2025).

Certified: AI that raises the bar for everyone

As AI becomes more integrated into healthcare, regulatory standards are rising to meet the challenge. Imaging AI must now meet stricter requirements to earn approval, proving it can be trusted not only to detect abnormalities, but to actively support diagnostic decisions.

icobrain aria is leading that shift. It’s the first FDA-cleared CADe/CADx device in neuroradiology, setting a new benchmark for quality and validation of AI imaging software.

"With icobrain aria, we’re leading the way in AI-driven neuroradiology. As the first FDA-cleared CADe/CADx device in the field, it sets a new benchmark for AI you can trust."

And let’s be clear: AI doesn’t replace clinicians. It supports their work. The final decision on diagnosis and treatment always remains with the experts who know their patients best.

Privacy & Security: Protecting patient data

In medical AI, trust extends beyond performance. It includes how patient data is protected. At icometrix, privacy and security are built into every part of our development process. We comply with international standards such as GDPR, HIPAA, are ISO 27001:2022 certified, and any imaging data used to train our models is carefully reviewed, handled under strict safeguards, and de-identified to the extent possible.

Security is not static. We continuously evaluate and strengthen our systems to stay ahead of evolving threats. As part of this, we maintain a responsible vulnerability reporting program that invites ethical hackers and researchers to flag potential risks - an open collaboration we proudly acknowledge in our public Hall of Fame.

Sustainability: Smart tech, responsible choices

Healthcare is resource-intensive, and AI should help, not add to the burden. We partner with Amazon Web Services (AWS), which is committed to achieving net-zero carbon emissions by 2040 and powering its operations with 100% renewable energy by 2025. Our cloud-based solutions leverage AWS's energy-efficient infrastructure, ensuring that our AI is as sustainable as it is powerful.

AI should improve healthcare without compromising the health of our planet. That’s a principle we stand by.

My personal view

As CEO of icometrix, I feel a personal responsibility for the impact of our solutions and the field of imaging AI as a whole.

I feel a responsibility to convince even the most conservative clinicians that AI is not a shortcut or a gimmick, but a trustworthy and reliable support. Something that helps them make better decisions, more confidently, for the people they care for. A second pair of eyes, which are great at quantifying change.

I feel a responsibility to the patients. I’ve seen what neurological disease can do. I know people, friends, and families who have lived through that uncertainty, who have waited for answers, for clarity. If we can help shorten that wait or bring more confidence to those moments, we should.

And I feel a responsibility to the planet our children and grandchildren will inherit. The choices we make today, how we build, how we run our infrastructure, how mindful we are of our impact, will shape the world our children and grandchildren grow up in. That matters to me deeply.

AI is about people, for people.

— Wim Van Hecke, CEO, icometrix