Human-AI coevolution

Human-AI coevolution, defined as a process in which humans and AI algorithms continuouslyinfluence each other, increasingly characterises our society, but is understudied in artificialintelligence and complexity science literature. Recommender systems and assistants play aprominent role in human-AI coevolution, as they permeate many facets of daily life and influencehuman choices through online platforms. The interaction between users and AI results in apotentially endless feedback loop, wherein users’ choices generate data to train AI models,which, in turn, shape subsequent user preferences. This human-AI feedback loop has peculiarcharacteristics compared to traditional human-machine interaction and gives rise to complexand often “unintended” systemic outcomes. This paper introduces human-AI coevolution as thecornerstone for a new field of study at the intersection between AI and complexity science focusedon the theoretical, empirical, and mathematical investigation of the human-AI feedback loop. Indoing so, we: (i) outline the pros and cons of existing methodologies and highlight shortcomingsand potential ways for capturing feedback loop mechanisms; (ii) propose a reflection at theintersection between complexity science, AI and society; (iii) provide real-world examples fordifferent human-AI ecosystems; and (iv) illustrate challenges to the creation of such a field ofstudy, conceptualising them at increasing levels of abstraction, i.e., scientific, legal and socio-political.

‍

‍