Huawei digital twin's window alert comes from a sensor, not AI

AI system

System: Nanjing Building 1:1 'living' digital twin on Huawei Cloud: 80-plus IoT sensor sets in 17 categories, 170-plus cameras, and 20-plus (Knews, Jiefang Daily) or 50-plus (Eastday) AI algorithms for smoking, crowd-flow and parking detection
Model: Not assessed
Release: Not assessed
Application developer: Huawei (Huawei Cloud, big data, AI, edge computing) with Huangpu District government under the Shanghai City Operations Management Center
Role: Not assessed
Configuration: Not assessed
Inputs and outputs: Not assessed
Autonomy: Not assessed
Workflow: Window-mounted contact sensors reporting to the building's digital twin. An alert pushed to a security guard's smart wristband. A guard on site who closes the window.
Human oversight: Not assessed
Relation to AI: For the window alert the sources document a window-mounted sensor whose signal passes through the Huawei Cloud digital twin to a guard's wristband. Knews and Jiefang Daily both say so, and neither attributes the detection to an AI algorithm. The 20- to 50-plus AI algorithms described run on cameras for smoking, crowd flow and bike parking. The one smoking alert shown at the launch was simulated by a Huawei engineer. The Monitor's statement that an AI system “detected a high-risk behavior (open window)” is therefore unsupported by its own sources.
AI contribution: For the window alert the sources document a window-mounted sensor whose signal passes through the Huawei Cloud digital twin to a guard's wristband. Knews and Jiefang Daily both say so, and neither attributes the detection to an AI algorithm. The 20- to 50-plus AI algorithms described run on cameras for smoking, crowd flow and bike parking. The one smoking alert shown at the launch was simulated by a Huawei engineer. The Monitor's statement that an AI system “detected a high-risk behavior (open window)” is therefore unsupported by its own sources.

Benefits

Outcome identifiers: Not assessed
Benefit: 995 door/window alerts from January to May 2021
Category: Increased Product Safety
Status: Not assessed
Beneficiary groups: Society. Organizations.
Beneficiaries: Not assessed
Reach: Not assessed
Distribution and access: Not assessed
Duration: Not assessed
Expected?: Yes
Costs and harms: Not assessed

Measurement

Studies: Not assessed
Study design: Not assessed
Measurement: Not assessed
Setting and sample: Not assessed
Measures: 995 door/window alerts from January to May 2021
Comparison: Not assessed
Results: 995 door/window alerts from January to May 2021
Additional value from AI: Not assessed
Limitations: Not assessed
Denominators: Not assessed
Materials for replication: Not assessed

Money

Kind of value: Not assessed
Amount: Not assessed
Currency and prices: Not assessed
Period and unit: Not assessed
Realization: Not assessed
Recipients and payers: Not assessed
Gross or net: Not assessed
Included costs: Not assessed
Comparison: Not assessed
Method: Not assessed
Return and payback: Not assessed

People and places

Participants: Developer: Huawei (Huawei Cloud, big data, AI, edge computing) with Huangpu District government under the Shanghai City Operations Management Center. Operator: Huangpu District City Operations Center and the Nanjing Building's operator (the Huawei flagship store).
Jurisdictions: Benefit: Nanjing Building, Nanjing Road East at Henan Road Central, Huangpu District, Shanghai, China. Developer: China. Operator: China.
Ownership and control: Not assessed
Dependencies: Not assessed
Cross-border connections: Not assessed
Geopolitical implications: Not assessed

Follow-up

Conditions for replication: A real, not visitor-triggered, window opening: Jiefang Daily reports that over 80% of 995 logged door/window alerts were visitors trying the devices. An identifiable risk to pedestrians below, which no source documents.
Replication attempts: 2 follow-up entries. Benefit recurred: 1. Outcome unresolved: 1.
Later findings: 7 May 2021. Jiefang Daily reported that from January to early May 2021 the digital twin had logged 995 door/window alerts, 439 smoking alerts and 55 fire-lane obstruction alerts, while an official said more than 80% of the handled events were triggered by visitors trying the IoT devices.
Availability: Not assessed
Institutional decision: Not assessed
Grounds: Not assessed
Action and result: Not assessed
Success criterion: Not assessed

Evidence

Source numbers: AIBD-0324-source-0001–0008
Citations: Numbered references below.
Language: Not assessed
Source relationships: The CAC page reposts news about the building. Smartcity.team compiles launch coverage. Neither is a regulatory finding or an independent study of the window incident. [6] [7]
Original evidence: The CAC page reposts news about the building. Smartcity.team compiles launch coverage. Neither is a regulatory finding or an independent study of the window incident. [6] [7]
Access and versions: Not assessed
Supported findings: Not assessed
Evidence links: Not assessed
Evidence strength: Not assessed
Missing evidence: Not assessed

Grades

Assessed benefits: Society. Organizations.
Depth: Not assessed
Reasons: Society: No pedestrian was shown at risk and no falling object occurred; the only documented trigger was a reporter's test, so no outcome is documented. Organizations: Jiefang Daily reports 995 door/window alerts handled in January to May 2021. It says that over 80% were visitors trying the devices and that real events were far fewer. No efficiency measure is given.
Observed or intended: Not assessed
Rubric: Not assessed
Assessor and date: Not assessed
Alternative assessments: Not assessed
Fill: Independently reported

Entry details

Kind: Not assessed
Related entries: Not assessed
Field: Urban governance and building safety. Smart-building monitoring under Shanghai's 'one network to manage all'.
Purpose: The digital twin was built to monitor the building's “vital signs” (vibration, smoke, lifts, crowd flow, open windows) and push alerts to staff. Window opening was named at the launch as a falling-object risk the system flags.
Dates: Date basis unassessed: 2021 (precision unassessed). Updated: 25 October 2026.
Stage: Not assessed
Version: SHA-256: 8669bbed8691…

References

Other identifiers: oecd_aim: 2021-02-25-adca.
Category mappings: Not assessed
Brief: Not assessed
Revisions: 7 revisions. Latest: 25 October 2026.

Sources

Registry label Independent author

OECD AI Incidents and Hazards Monitor, record 2021-02-25-adca, 'Shanghai Deploys AI Digital Twin for Real-Time Urban Safety Management' (AI-generated summary).
News report Independent author

Eastday (Xue Ningwei), 25 February 2021, '改变从一栋楼开始！上海"一网统管"探索城市最小管理单元数字治理', via Sohu.
News report Independent author

Knews / Shanghai Media Group, 25 February 2021, '南京大楼，如何实现"数字孪生"？', via NetEase (163.com).
News report Independent author

Jiefang Daily, 7 May 2021, '传感器如"五官"，南京大楼在数字世界"活"了', via Guangming Online (GMW).
News report Independent author

Banyuetan / Xinhua (Zhou Lin), 6 July 2021, '"数字孪生城市"来了：城市治理可望告别"盲人摸象"', via The Paper.
News report Independent author Context or contrary evidence

Cyberspace Administration of China, 4 March 2021, repost of '改变从一栋楼开始！上海"一网统管"探索城市最小管理单元数字治理' from the 'Cyberspace Shanghai' WeChat account.

Hardware and platform context; a repeated window-alert account. Source credit and 实时体验: 大楼也有数字生命体征.

A government-hosted news repost, not a regulatory instrument.
News report Independent author Context or contrary evidence

smartcity.team, 1 March 2021, '上海：城市最小管理单元数字治理成果发布，数字治理助力一网统管' (case write-up of the 25 February 2021 launch).

Compiled launch coverage and a repeated incident description. Byline 本站综合编辑; launch and window passages.

Secondary compilation, not independent benefit verification.
Press release Developer

zhiding.cn (big-data channel), 25 February 2021, '上海"一网统管"发布城市最小管理单元数字治理成果', marked '来源：业界供稿' (industry-supplied copy).