A featured contribution from Leadership Perspectives, a curated forum for enterprise technology leaders, nominated by our subscribers and vetted by the CIOApplications Editorial Board.
Genmab
Advancing Precision Medicine through Responsible AI Leadership
Translational Data Leadership: Making Oncology Personal Through Data-Driven Insights
As VP and Global Head of Translational Data Sciences, I lead a team dedicated to transforming complex biological and clinical data into meaningful insights that help us understand how our oncology medicines work—and for whom.
m. Our guiding principle is clear: the right drug for the right patient, at the right dose and time. We use advanced analytics, AI, and machine learning to extract insights from a wide range of patient data—from molecular and cellular markers to medical imaging—building a comprehensive view of each individual’s characteristics.
These insights help us characterize drug mechanisms, uncover disease heterogeneity, inform combination strategies, and predict treatment responses. Ultimately, everything we do supports one goal: enabling more personalized and effective treatment decisions for cancer patients.
Underrated AI Challenges: Context and Culture
One of the most underrated bottlenecks in AI adoption—particularly in life sciences— is the complexity of working with unstructured and semi-structured data.
Clinical notes, pathology reports, radiology reads, patient registries, and insurance claims are often inconsistent, unstandardized, or lack derived variables. Cleaning and structuring this data is time-intensive and demands deep domain expertise to extract meaningful, context-driven insights.
Another challenge lies in the disconnect between a model’s output and its clinical or biological relevance. AI can generate thousands of features from medical images, but many lack interpretability.
Lastly, organizations often underestimate the cultural and operational shifts needed for AI to take root. It’s not just about the model— building trust, reshaping collaboration, and evolving workflows to incorporate and act on AIdriven insights.
Dark Data in AI: Fuse Modalities to Reveal Hidden Insight
Organizations need to recognize that critical insights often lie within rarely analyzed data—what we call the “dark matter” of AI. This includes patient-reported outcomes, adverse event reports, physician notes, wearable device data, and lesser-used imaging modalities. While messy and highdimensional, this data can provide essential context that structured sources miss.
In oncology, it’s not enough to generate data—we need to connect it to biology, patient context, and real-world decisions to make AI truly impactful
To unlock its value, teams must invest in the infrastructure and talent to handle diverse data types. That means building multi-modal pipelines capable of harmonizing everything from free text to pixel-level image data. For example, radio mic or histology features become exponentially more meaningful when integrated with clinical or molecular data. This fusion can uncover drug response or resistance mechanisms that would otherwise remain hidden.
AI Ethics Oversight: New Structures Must Meet Technical and Sustainability Demands
Traditional ethics committees are rooted in clinical protocols and human subject protection. But AI presents different challenges: algorithmic bias, data privacy, explainability, and model misalignment, among others. These issues require dedicated AI ethics boards with cross-disciplinary expertise, bringing together data scientists, ethicists, clinicians, patient advocates, legal and regulatory experts, pharmacovigilance professionals, and sustainability specialists who understand AI systems' technical depth and real-world consequences.
Crucially, sustainability must be a foundational part of AI governance. Large-scale models consume enormous energy and carry a tangible carbon footprint. Any credible AI ethics framework should include practices that optimize model efficiency and ensure environmentally responsible infrastructure.
Using AI for Transformation: Think Bigger but Stay Grounded
Visionary leaders should focus on AI applications that unlock new capabilities, not just automating existing ones. Avoid the temptation to apply AI indiscriminately. Instead, prioritize areas with measurable efficiency gains and strong validation frameworks, especially where hallucinations or model drift could undermine trust and safety.
Think big, but execute deliberately. The most transformative AI emerges at the intersection of disciplines. Build cross-functional teams that blend data science, ethics, domain expertise, and design thinking. And always keep human oversight in the loop—particularly in sensitive fields like biopharma, where nuance, context, and precision aren’t optional, and where patient impact is real and immediate.
