When proteins speak and AI listens
What happens when biology becomes a data-rich science, and artificial intelligence is ready to listen?
At the Hi! PARIS Summer School 2025, Jean-Philippe Vert, Co-founder and CEO of Bioptimus, and a leading researcher at the intersection of AI and medicine, shared how machine learning is transforming the foundations of biomedical research. With a career spanning cancer research, computational biology, and deep learning, Vert outlined a simple yet powerful thesis: we are entering a golden age for AI in biology, where breakthroughs are no longer hypothetical, they are happening now.
Key Takeaways
- Modern biology generates massive volumes of data, DNA sequences, protein structures, tissue images, at multiple scales.
- AI architectures are now powerful enough to learn from and integrate this complexity.
- AI systems like AlphaFold have already solved scientific problems that remained untouched for decades, revolutionizing protein structure prediction.
- Biology is fundamentally multiscale, requiring new AI models capable of moving between molecules, cells, and organisms.
- The next frontier includes in silico simulations of disease, virtual drug testing, and predictive models of molecular interactions.
Turning scientific complexity into computational power
The case of AlphaFold, an AI system developed by DeepMind to predict the 3D structure of proteins, has become emblematic of the power of machine learning in biology. A problem that resisted decades of effort was resolved with an architecture trained on massive datasets. The scientific community responded in kind: AlphaFold is now in widespread use across labs, and its impact was formally recognized by the 2024 Nobel Prize in Chemistry.
For Vert, the importance of this moment lies not just in the result, but in what it signals: AI is no longer a supportive tool for science, it is a driver of discovery.
When biology demands more than language models can offer
Unlike language, which often exists in linear, single-scale formats, biology operates at multiple levels of organization. DNA sequences are just the beginning. Scientists now collect information at the level of proteins, cells, tissues, and patients. Each level has its own logic, but they are interdependent, and understanding disease, or designing new treatments, requires seeing across them. This presents both a challenge and an opportunity for AI. Current architectures, designed largely for language or vision tasks, need to evolve. New models must integrate heterogeneous data and learn how biological systems behave across scales. In Vert’s view, this is not just an application of AI, it is a new frontier for its development.
Jean-Philippe Vert at the Hi! PARIS Summer School 2025
In Silico Medicine: The Future Has Begun
Looking ahead, Vert believes the next generation of AI tools will go beyond structural prediction. We will see models capable of:
- Simulating how diseases progress over time
- Predicting how molecules interact inside the body
- Testing treatments virtually before human trials begin
These capabilities open the door to in silico experimentation, a way to accelerate science without the same resource or ethical constraints. For drug discovery, personalized medicine, and clinical diagnostics, the implications are enormous.
Advice for those ready to explore the frontier
Having worked within both academic and institutional circles, including ADIA and CFM, the speaker underscored a paradox. The strongest systematic investment groups often rely on deeply internal knowledge systems. While this drives focus and cohesion, it can also create blind spots.
Some ideas are dismissed too quickly; others are adopted too casually. “Seeing nails everywhere once you have a hammer” is a risk familiar to both researchers and practitioners. The key is pragmatic humility, a mindset that treats academic literature as a resource to be tested, not a script to be followed
A new frontier for students and researchers
For students and researchers considering a career in this space, JP Vert’s message is clear: don’t wait. “Just go for it,” he said. The field is moving fast, and interdisciplinary talent is in high demand.
Whether your background is in biology, AI, or engineering, the key is collaboration. Biologists need to become fluent in data science. Engineers need to understand biological systems. The tools exist, and so does the data. What’s needed now is a new generation of scientists who can work across domains, and think across scales.