The Concepts Matter. Or: Reflections on “Imaginations of Autonomy”

May 23, 2024

In the context of soccer-playing robots, technological advances in robotics and artificial intelligence are mostly perceived as exciting and fun. Whenever I tell people about my research project, they tend to be very curious to learn more about the state of the technology. How well can the robots play soccer? Do they use artificial intelligence? How do they make decisions? And are they going to win the match against humans in 2050?

When it comes to robot soccer, many people perceive the use of the latest technology to allow robots to score more goals - or get fewer time penalties - as desirable. However, similar technological approaches in fields such as object detection or data- and calculation-based decision-making are also used in highly destructive ways.

I am currently attending the conference “Imaginations of Autonomy: On Humans, AI-based Weapon Systems and Responsibility at Machine Speed” at Paderborn University, organized in the context of the competence network “Meaningful Human Control”.

On the first day of the event, during the pre-conference workshop, I learned a lot about the global use of autonomous weapon systems, their geopolitical relevance, and the power structures they reproduce. In this blog post, I want to discuss one aspect in greater detail and relate it to my research, namely that concepts materialize in technologies.

One of the concepts discussed most extensively in the pre-conference workshop was “autonomy.” Similarly to other concepts often drawn upon in technology development – intelligence, curiosity, companionship, etc. – “autonomy” can be conceptualized in different ways. Different conceptualizations can, in turn, lead to various implementations.

In my research, for example, the ambivalence of this concept came up in a discussion about how strongly individual soccer robots should rely on communication with other robots to make their decisions. Should the software aim at gaining a sufficient understanding of the surroundings in each individual robot, or should it aim at using the information gathered by the sensors of all of the team’s robots? It is also worth discussing how the role of humans is conceptualized in robot autonomy. While during a robot soccer game, the rules of the game (which were set by humans) only allow for limited human intervention, the robots could never play soccer without quite a lot of preparatory work by humans.

The influence of concepts should thus not be overlooked in the ethical design of technologies. Discussing and critically scrutinizing the concepts used as guiding principles in technology development, particularly the ones whose meaning is often taken for granted, can help us better understand the technologies that have already started to shape our lives in numerous ways.