How can artificial intelligence (AI) be designed to meet the demands of a democratic society? In a new video, technology sociologist Nils Klowait, a research associate in the SFB/TRR 318’s public outreach project, addresses this question and explains the concept of co-construction. He highlights that technological systems should not only be developed by engineers or scientists but also in dialogue with various societal groups. This approach is central to the work of the TRR 318 "Constructing Explainability," which focuses on the transparent and comprehensible development of AI.
AI is increasingly permeating all aspects of life, from music recommendations on streaming platforms to critical decisions in healthcare and lending. But who designs these systems, and whose interests shape their development? Klowait emphasizes that AI, like any technology, is not neutral. It is influenced by the perspectives, biases, and values of those who create it. This is where the concept of co-construction comes in: by actively including diverse voices, more just and inclusive technology can emerge.
Researchers in TRR 318 are exploring ways to make AI systems more responsive and understandable. This includes "Co-Construction Workshops," where people from various social groups interact with modern AI systems and contribute their ideas about where, whether, and how, these systems should be applied. These workshops help develop technological solutions that address the needs of a diverse population.
A particularly striking example of the co-construction approach is the "Perpetual Prototype." Originally designed as a small robotic head for workshop participants, this model continuously evolved through participant feedback. Requests such as "I want to attach the robot to my keychain" or "It should fit on a pencil" led to new versions of the prototype. This open and dynamic development process demonstrates how democratic design principles can lead to unexpected yet meaningful technological solutions. The robot symbolizes the continuous negotiation process that AI systems should also undergo.
"Artificial intelligence is not just a technical tool; it's a reflection of the society that shapes it," explains Klowait. "If we are serious about creating a just and democratic society, we must ensure that AI systems are developed not just for people, but with them." The "Perpetual Prototype" illustrates how such a process can unfold, serving as inspiration for the development of AI systems that remain open to change and modification. Ultimately, the co-construction of technology is not just an idealistic approach but a practical necessity. By ensuring that diverse voices are included in the development process, we can build a future where AI systems are not only efficient but also just and inclusive.
The video is now available on the TRR YouTube channel.