China’s AI training-data push highlights why vertical datasets matter more than generic chat
SCMP Tech reported on 9 June 2026 that China’s data authority released a draft nationwide plan to expand the supply of high-quality AI training data. The report says the plan focuses on industry-specific datasets as developers face a looming shortage of useful data for next-generation models.
The Bubbll takeaway is simple: vertical AI wins when it understands the job. For retail, restaurant, hotel and residence workflows, the valuable data is not random conversation volume; it is product catalogs, booking states, guest preferences, support outcomes, consent records and staff feedback.
That makes data design a product feature. Bubbll should treat every customer interaction as a permissioned signal that can improve routing, personalization and service recovery while respecting privacy boundaries. Better datasets create better assistants; better governance keeps those assistants safe enough for real businesses.
Source: SCMP Tech
Image: “Artificial Intelligence & AI & Machine Learning” by mikemacmarketing, licensed under CC BY 2.0 via Wikimedia Commons. Attribution requested by source: Image via www.vpnsrus.com. Image source/license page: https://commons.wikimedia.org/wiki/File:Artificial_Intelligence_%26_AI_%26_Machine_Learning.jpg
