What happened
KR-Asia reports that on April 30, 2026, Xiaohongshu sent an internal letter to employees announcing the creation of Dots, described as a first-level AI department. According to KR-Asia and reporting from 36Kr, Dots is organised into four divisions: model R&D, infrastructure, engineering, and product, and it reports to Xiaohongshu president Ding Ling (alias Conan). KR-Asia reports that Dots was upgraded from the former Hi Lab (Humane Intelligence Lab) and that the company's key internal AI application, Diandian, has been placed under the new unit.
Editorial analysis - technical context
Industry-pattern observations: platforms that centre on user-generated content frequently face a tradeoff between automated AI features and perceived authenticity. After the emergence of ChatGPT in late 2022, many consumer apps experimented with model-driven assistants, recommender enhancements, and content synthesis, often introducing layers of moderation and human-in-the-loop controls to preserve community norms. For engineering teams, combining model R&D and infrastructure under a single organisation commonly aims to accelerate iteration cycles, reduce integration friction, and standardise deployment and safety tooling across product surfaces.
Industry context
Editorial analysis
Reporting places Xiaohongshu's change in the broader wave of companies moving from cautious experimentation to more formalised AI organisations. Observers covering comparable transitions note that elevating AI into a first-level unit often coincides with increased executive oversight and tighter cross-functional coordination, which can shorten time-to-market for AI features while raising governance questions about moderation, content provenance, and advertiser alignment.
What to watch
For practitioners and observers
track public product changes that place model outputs into feeds or search results, any updates about Diandian features, and signals from Xiaohongshu on moderation policy adjustments. Also watch whether Xiaohongshu publishes engineering or safety documentation for model deployment and how third-party partners or advertisers respond to changed AI behaviour.
Note: KR-Asia is the primary source for the reported organisational changes, and 36Kr is cited within KR-Asia for the upgrade from Hi Lab. Xiaohongshu has not been quoted directly in the reporting cited.
Key Points
- 1Xiaohongshu formalised AI with a new first-level unit, signalling operational escalation in model work and infrastructure integration.
- 2Combining model R&D, infrastructure, engineering, and product under one org typically speeds feature delivery but raises governance needs.
- 3Observers should monitor public product changes, moderation updates, and any published deployment or safety documentation from the company.
Scoring Rationale
This is a notable corporate organisational shift at a major consumer platform with implications for productisation of AI and governance. The change matters to practitioners but is not a frontier research or platform-level disruption.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
