News

What remains for physicists to do in the age of AI?
Giovanni Volpe
OPTIQUE BFC 2026
Date: 9 July 2026
Place: Palais des Congrès, Dijon, France

Giovanni Volpe gave the plenary presentation “What remains for physicists to do in the age of AI?” at OPTIQUE BFC 2026, the biennial congress of the Société Française d’Optique, held in Dijon, France, from 6 to 10 July 2026.

In recent years, the rapid growth of artificial intelligence, particularly deep learning, has transformed fields ranging from the natural sciences to technology. While deep learning can be regarded as a highly sophisticated form of curve fitting, the development of increasingly large and complex neural networks has led to remarkable advances, often surpassing expectations. As AI systems become ever more capable, this raises a fundamental question: what role remains for physicists and, more broadly, for humans?

A critical yet often overlooked aspect of current AI systems is their reliance on vast amounts of training data, much of which ultimately originates from humans. This raises important questions about what happens when human-generated data become scarce or when AI systems increasingly train on AI-generated content. Such feedback can lead models toward increasingly conventional or average outputs, potentially reducing originality and reliability.

In this context, physicists have a distinctive role to play. Their ability to develop fundamental understanding, identify symmetries and physical principles, design experiments, and formulate new questions can provide insights that cannot simply be extracted from existing datasets. Physics-informed approaches can also incorporate this knowledge directly into the architecture and training of AI systems.

The central message is that, in the age of AI, physicists should continue to be original, pursue fundamental questions, and search for the hidden principles governing the physical world.

Autonomous Robots for Intelligent Matter
Giovanni Volpe
Toward Intelligent Behavior in Macroscopic Active Matter
Time: 09:00
Date: 6 July 2026
Place: CECAM-HQ-EPFL, Lausanne, Switzerland

Giovanni Volpe gave the invited talk “Autonomous Robots for Intelligent Matter” at the CECAM workshop “Toward Intelligent Behavior in Macroscopic Active Matter”, held at CECAM-HQ-EPFL in Lausanne, Switzerland, from 6 to 8 July 2026.

Active matter provides a framework for understanding systems composed of self-driven units, ranging from microscopic particles to animal groups and robotic swarms. While traditional active-matter models often describe agents through relatively simple interaction rules, autonomous robotic systems introduce additional capabilities such as sensing, information processing, decision-making, memory, and adaptation.

In his talk, Giovanni discussed how autonomous robots can provide experimental platforms for exploring these emerging forms of intelligent matter. By combining self-propulsion and physical interactions with sensing, feedback, and learning, robotic agents can adapt their behaviour to their surroundings and coordinate with one another, providing a bridge between active-matter physics and embodied intelligence.

These systems offer opportunities to investigate how complex collective behaviours emerge from interactions between individual agents and their environment, while also providing new approaches for designing adaptive and programmable matter. More broadly, they illustrate the growing convergence between statistical physics, robotics, nonlinear dynamics, and machine learning in the study of intelligent active systems.

Microscale systems offer a unique opportunity to engineer machines whose function emerges from the interplay of geometry, interactions, and fluctuations. In this talk, I will present our work on smart machines at the microscale, combining nanofabrication, programmable interactions, and advanced optical methods to design and control colloidal micromechanisms and metamaterials.
I will first introduce how nanotechnology enables the realization of colloidal metamachines and microscopic mechanisms, where shape and mechanical constraints are engineered to produce targeted motion and response in fluid environments. I will then show how smart microscopy and optical manipulation—including high-resolution imaging, automated tracking, and light-based control—allow us to probe these machines in real time and quantify their dynamics. This approach enables precision measurements of effective interaction landscapes, including critical Casimir forces and their relation to fluctuation-induced forces known from QED Casimir physics.

In recent years, the rapid growth of artificial intelligence, particularly deep learning, has transformed fields from natural sciences to technology. While deep learning is often viewed as a glorified form of curve fitting, its advancement to multi-layered, deep neural networks has resulted in unprecedented performance improvements, often surprising experts. As AI models grow larger and more complex, many wonder whether AI will eventually take over the world and what role remains for physicists and, more broadly, humans.

A critical, yet underappreciated fact is that these AI systems rely heavily on vast amounts of training data, most of which are generated and annotated by humans. This dependency raises an intriguing issue: what happens when human-generated data is no longer available, or when AI begins to train on AI-generated data? The phenomenon of AI poisoning, where the quality of AI outputs declines due to self-referencing, demonstrates the limitations of current AI models. For example, in image recognition tasks, such as those involving the MNIST dataset, AI tends to gravitate towards ‘safe’ or average outputs, diminishing originality and accuracy.

In this context, the unique role of humans becomes clear. Physicists, with their capacity for originality, deep understanding of physical phenomena, and the ability to exploit fundamental symmetries in nature, bring invaluable perspectives to the development of AI. By incorporating physics-informed training architectures and embracing the human drive for meaning and discovery, we can guide the future of AI in truly innovative directions. The message is clear: physicists must remain original, pursue their passions, and continue searching for the hidden laws that govern the world and society.