https://aifaultline.com/authors/mingyi-ma/New papers by Mingyi Ma in the AI Fault Line guide, and news naming them. https://aifaultline.com/research/p/title-1b8acf883a6f873f/https://aifaultline.com/research/p/title-1b8acf883a6f873f/Mon, 25 May 2026 00:00:00 +0000Using 5 million Beijing job postings (2018-2024) and an LLM-derived neighborhood-level GenAI Exposure Index, the authors show exposure concentrates in core districts and that since 2023 high-exposure neighborhoods have suffered wage stagnation even while attracting high-skilled workers — a "high-skill trap" driven by task de-skilling and labor-market crowding, causally identified with a diff-in-diff around ChatGPT's release. This is a direct empirical challenge to skill-biased technological change as the operative model for GenAI, and one of the first studies of intra-urban (rather than cross-occupation) AI inequality.