AI algorithms and how they interact with Society

Artificial Intelligence (AI) is here to stay and, with the recent buzz about ChatGPT, has become omnipresent in public discourse. Recent polls suggest that many people have mixed feelings about AI algorithms and how they influence their lives, but at the same time, also little knowledge about what AI is and how exactly it is used. In this talk, I will first give a general introduction into AI algorithms and how they differ from traditional, non-AI algorithms. Then, I will show some insights from recent research projects which look into the interaction of AI and society. In the ReNewRS projects, we analyze polarization effects by news recommenders and how they can be mitigated. In the KarekoKI project, we investigate whether algorithmic pricing engines can learn to cooperate and coordinate prices, and whether or not they violate antitrust legislation. Along the observations from those projects, I will get back to the difference between traditional and AI algorithms and discuss why developing AI algorithms needs a different focus and skill set than the development of traditional algorithms.