What should AI do next? Symposium with inaugural lectures of new foreign members

Recently AI has moved from a small corner of computer science into the full spotlight, grabbing almost daily headlines with new achievements and recognition, even with two Nobel prizes this year. Although the impact of AI is already very high and widespread, there are clear limitations and there is much application potential that has not been addressed yet. This symposium brings together three distinguished scientists in the fields of AI, Complex Systems Science, and Evolutionary Biology. Based on their excellence and scientific leadership they are being incorporated as foreign members in the section for Natural Sciences of the Academy. At this symposium they will talk about the breakthroughs they themselves have achieved in their own discipline and reflect on the potential they see for future AI in the advancement of science and society.

The full circle of AI: from logic to learning and back again

Frank van Harmelen (Free University Amsterdam)

The fundamental question of Artificial Intelligence (AI) is: "is thinking simply computing?", and more precisely "which type of computing?". While most AI researchers hypothesise that the answer to the first question is "yes", the second question is much more controversial, and opinions have varied greatly in the young history of AI. Different schools have formed around the "physical symbol systems hypothesis" of Newell and Simon (emphasising the importance of deduction), the "Embodied Cognition hypothesis" of Rodney Brooks (emphasising the role of the physical world) and the "Distributive Semantics hypothesis" (emphasising the importance of induction) by J.R. Firth. Some in AI (and me among them) would argue that after decades of divergence, the time has now come for a period of convergence, where we visit the lessons of the past with an eye to an integrative computational theory of intelligence.

The adventure of the new: from Laplace to AI

Vittorio Loreto (La Sapienza University)

New experiences continually shape our lives - whether through meeting new people, discovering new ideas, or adopting new technologies. While some of these experiences arise by chance, many are influenced by prior novelties, creating a web of correlation and causation. Throughout history, new developments have presented opportunities and challenges, often challenging our innate desire to predict the future. Understanding how novelties emerge and influence our actions is crucial. The intuitive idea that one novelty often leads to another is formally expressed in the concept of the "adjacent possible." This concept encompasses all ideas, structures, and artefacts, just one step beyond our current understanding. I have contributed to formalizing this concept within a solid mathematical framework. After presenting a historical perspective on how these phenomena have been addressed, I will explore how complexity science connects social and human sciences—focusing on areas such as the rise of language conventions, consensus-building, evolutionary writing, and polarization. Finally, I will discuss how the notion of the new combined with AI technologies paves the way for a new form of human creativity.

A four-billion-year journey across the evolutionary landcape: the major transitions in evolution

Eörs Szathmáry (Eötvös University, Budapest)

There is no theoretical reason to expect evolutionary lineages to increase in complexity with time, and no empirical evidence that they do so. Nevertheless, eukaryotic cells are more complex than prokaryotic ones, animals and plants are more complex than protists, and so on. This increase in complexity may have been achieved as a result of a series of major evolutionary transitions. These involved changes in the way information is stored and transmitted and the formation of higher-level evolutionary units from lower-level ones. I shall give a personal account of this journey, looking at insights and new experiments. Some of the transitions, including the origin of natural language, will be highlighted.

Maynard Smith and I wrote in a popular book in 1999: "And the latest transition that is taking place in our lifetimes is the proliferation of electronic ways of storing and transmitting information... Will our descendants spend most of their lives in virtual reality? Is there a symbiosis between genetic and electronic information storage? Will electronic devices be able to self-replicate to replace the primitive life forms that created them? We do not know." On this note, I shall briefly consider what good and bad AI can bring us as a contributor to evolutionary understanding and, possibly, as a novel evolutionary system as such.

About the speakers

Frank van Harmelen is Professor of Computer Science. Knowledge Representation and Reasoning Group. Network Institute, Free Unversity Amsterdam (VUA). Member Dutch Academy of Arts and Sciences (KNAW).

Vittorio Loreto is Professor of Physics. Department of Physics, La Sapienza University, Rome. Sony Computer Science Laboratory, Rome.

Eörs Szathmáry is Professor of Biology. Centre for Ecological Research of the Hungarian Academy of Sciences. Institute of Biology, Eötvös University, Budapest. Parmenides Center for the Conceptual Foundations of Science, Pocking, Germany. Member Hungarian Academy of Sciences (MTA).

Luc De Raedt is Professor of Computer Science. Department of Computer Science, KU Leuven, Belgium. Member, Class of Natural Sciences, Royal Flemish Academy of Belgium.

Luc Steels is Professor Emeritus of Computer Science. Artificial Intelligence Laboratory, Vrije Universiteit Brussel (VUB), Brussels. Chair Section Natural Sciences, Royal Flemish Academy of Belgium.

Symposium organised in cooperation with the Evolutionary Linguistics Association (ELA).