3 Questions: Beyond data-driven aesthetics

June 30, 2026 - 00:41

MIT Architecture alumnus and researcher Alexandros Haridis is challenging how we think about beauty and machines. His new exhibition, "Beyond Data-Driven Aesthetics," currently on display at the Keller Gallery, does not simply show pretty pictures made by a computer. Instead, it digs into the deeper history of computational aesthetics. Haridis translates complex algorithms, machine learning systems, and old theories about what makes something beautiful into physical installations and interactive visualizations.

The work asks a fundamental question: what happens when we stop letting data alone decide what looks good? Haridis argues that modern AI often relies on massive datasets to mimic style, but this approach ignores centuries of philosophical thought on aesthetic judgment. His exhibition pulls from early cybernetics and 20th-century information theory to show that the quest for a "beautiful" algorithm is much older than the current AI boom. Visitors can engage with pieces that make abstract mathematical rules tangible, turning code into something you can walk around and touch. The goal is not to reject technology, but to broaden the conversation about how we define creativity and taste in an age of generative models.