What happened
The South China Morning Post reports that Zhu Shuhan, a 13-year-old junior-secondary student in Shanghai, created an AI app called Zai Chang, or “On Site,” to help parents supervise children during homework. The app uses a parent’s uploaded photo to create a virtual parent that issues voice reminders when its camera detects mobile-device use or that a child has left their seat. It also generates a concentration report after homework is completed.
The South China Morning Post reports that Zhu Shuhan began using AI only in 2025 and had no programming background when she developed Zai Chang. She is identified as a Grade Two student at a junior secondary school in Shanghai. The report presents the app as a tool for parents who want help monitoring children’s study habits, rather than as a general-purpose education platform.
According to the South China Morning Post, the app was entered in a national AI product marketing competition held in Hangzhou from August 21 through August 23. More than 50 participants competed. The contest reportedly did not have technical judges; instead, rankings were determined by the income generated by participants’ AI-created apps during the three-day period.
The report says Zai Chang received 90 orders and generated total sales of 18,000 yuan, or about US$2,600. Those reported sales placed Shuhan third and earned her a 5,000-yuan prize. The source does not identify the buyers, the price structure beyond the total sales, or whether the orders represented subscriptions, one-time purchases or another form of transaction.
The app’s reported workflow combines an uploaded image of a parent with camera-based monitoring. Once a parent uploads a photo, the system assumes the role of a virtual parent. If the camera detects that a child is using a mobile device or has left their seat, the virtual parent displayed by the app gives a voice reminder. After the homework session, the app generates what the report calls a “concentration report.” The source does not describe the underlying model, the hardware used, or how detection decisions are made.
Why it matters
The report provides a concrete example of a young person using AI tools to build and sell a functional consumer application without programming experience. It also shows how an AI product contest evaluated commercial results rather than technical judging. However, the account describes one competition and does not establish that similar products are broadly viable, effective or safe.
The case is significant as a specific example of AI-assisted product creation becoming accessible to someone without formal programming skills. The South China Morning Post’s account suggests that the barrier in this competition was not only technical implementation but also identifying a use case, assembling an AI-generated app and finding paying users. That is evidence of one successful project, not proof that programming expertise is no longer needed for reliable software development.
The contest also connects AI development with immediate market testing. Because the reported rankings were based on income during a fixed three-day period, the event rewarded an app’s ability to attract orders rather than its technical novelty or independently measured educational benefit. That makes the result useful as a snapshot of commercial interest, while limiting what can be inferred about quality, safety or long-term adoption.
The product targets a familiar household problem: parents’ difficulty supervising homework. A voice reminder and post-session report could make an AI system feel more tangible to families than a general chatbot. At the same time, the source does not show whether the app improves concentration, homework completion or learning outcomes. The report gives no comparison with ordinary parental supervision or other study tools.
The app also raises practical questions because it appears to monitor children through a camera and uses a parent’s image to represent an automated supervisor. The South China Morning Post does not say whether video or images are stored, where data is processed, how long information is retained, whether parents provide consent for all people captured by the camera, or how the system handles mistaken detections. Those omissions matter more than the sales figure for assessing whether the product is suitable for routine use.
What to watch next
The source does not independently establish the app’s technical performance, underlying AI system, privacy safeguards or continuing availability. Further reporting should verify the sales, determine whether the 18,000 yuan represents gross revenue, and examine how the app handles images, camera data, voice reminders and children’s information.
The first issue to watch is verification. The South China Morning Post attributes the sales and competition result to its reporting, but the supplied article does not include independent payment records, an app listing, an interview with buyers or technical documentation. Confirmation of the 90 orders and 18,000-yuan total would clarify whether the figure represents completed transactions, commitments, gross sales or another contest metric.
The next question is whether Zai Chang continued operating after the August 21-23 competition. The source does not provide an availability date, download count, repeat-purchase data, refund information or evidence of use outside the contest. Follow-up reporting could establish whether the app was a short-lived competition entry or a product with sustained demand.
Technical performance also remains unknown. The article does not identify the camera system, AI model or detection thresholds, and it does not report false alarms, missed detections or performance in different lighting and household settings. A meaningful assessment would need to examine how often the app incorrectly concludes that a child is distracted or out of their seat, and whether parents can review or override those decisions.
Privacy and child-safety practices deserve particular scrutiny. The source does not explain how the uploaded parent photo, camera footage, voice data or concentration reports are protected. It also does not say whether the app collects information about children beyond the immediate session. Any continued deployment would make those details important for families, schools and regulators evaluating AI tools used around minors.