Source record

Published by arXiv cs.HC (T1 source). The original is at https://arxiv.org/abs/2609.38183.

Pipeline notes

The summary and note below are generated by the signal pipeline — they are Beyond Desk’s reading, not quotations from the source.

Empirical study of 73 promotional websites built by graduate students using the AI coding tool Lovable finds that designs cluster into a small set of visual patterns, indicating collective homogenization despite subjective feelings of originality. Authors name this 'authored ignorance' and propose measurement plus design interventions.

Quantifies a specific failure mode of single-tool AI generation — visual convergence across users — with embedding data, process recordings, and survey evidence, plus concrete design implications for originality feedback.

Cited by

No citations on record.