New paper reveals 4 aesthetic biases that shape AI research

A new paper by Angelina Chen, Rick Sullivan, and Raffaele F. Ciriello, accepted at the European Conference on Information Systems (ECIS 2026), argues that Information Systems research is deeply influenced by implicit aesthetic assumptions—what researchers consider beautiful, worthy, or important. Through a hermeneutic literature analysis, the authors surface four distinct aesthetic perspectives: aesthetics as imitation (valuing faithful representation), sensory experience (focus on human perception), world-making (how systems co-create realities), and political doing (how power shapes what is seen). The study demonstrates that these perspectives act as epistemic infrastructure, conditioning what questions get asked and what findings are deemed legitimate.

Applying their framework to two hot topics—algorithmic management and digitally mediated intimacy—the authors show how each aesthetic lens opens different research questions and exposes blind spots in dominant approaches. For example, a 'political doing' lens on algorithmic management reveals hidden labor power dynamics, while a 'sensory experience' perspective on digital intimacy highlights neglected somatic dimensions. The paper provides a much-needed vocabulary for AI and HCI scholars to reflect on their own theoretical commitments, potentially reshaping how future studies frame problems and contributions. For AI professionals, this work underscores that technical fields are not value-neutral—aesthetic choices influence everything from model design to deployment ethics.