From Score-Driven to Value-Sharing: Understanding Chinese Family Use of AI to Support Decision Making of College Applications

TL;DR

This study investigates how AI tools like Quark GaoKao are used in China's high-stakes GaoKao college admissions to support decision-making among students, parents, and experts. Through 32 qualitative interviews, it finds that AI use is largely parent-led and score-centric, with limited student engagement and insufficient attention to long-term career goals. The work identifies challenges such as data localization gaps, information asymmetries, marketing-driven consultant practices, and regional policy complexities, and offers family-centered design insights to improve collaboration and equity. The findings highlight socio-technical barriers—like varying literacy, internet access, and social capital—that shape who benefits from AI-assisted college planning, underscoring the need for accountable, transparent, and accessible AI tools for families.

Abstract

This study investigates how 18-year-old students, parents, and experts in China utilize artificial intelligence (AI) tools to support decision-making in college applications during college entrance exam -- a highly competitive, score-driven, annual national exam. Through 32 interviews, we examine the use of Quark GaoKao, an AI tool that generates college application lists and acceptance probabilities based on exam scores, historical data, preferred locations, etc. Our findings show that AI tools are predominantly used by parents with limited involvement from students, and often focus on immediate exam results, failing to address long-term career goals. We also identify challenges such as misleading AI recommendations, and irresponsible use of AI by third-party consultant agencies. Finally, we offer design insights to better support multi-stakeholders' decision-making in families, especially in the Chinese context, and discuss how emerging AI tools create barriers for families with fewer resources.

Figures (6)

• Figure 1: An example of using Quark GaoKao for a user from Henan Province with a GaoKao score of 600, and ranking position of No. 23638. After the user inputs their scores and PA (Step 0), they can set filter and sorting preferences (Step 1) to view an AI-generated list of colleges with admission probabilities (Step 2). Users may refine the AI-generated college list through features that helps validating and narrowing down (Step 3). Then, users may output the application form to be submitted to the government system. Highlighted sections: ① Filter options for colleges based on the probability of admission: 16 reach schools are hard to get into (in red), 46 target schools where users' chances are good (in blue), and 447 safety schools that are easier to get into (in green). ② Four sorting options for colleges: probability of admission descending/ascending, planned admission number descending/ascending. ③ Name, code, category, and overall probability of admission to the selected colleges. ④ Cutoff scores and cutoff ranking positions in the past years for the selected colleges. ⑤ AI-generated FAQs regarding the selected colleges. ⑥ Admission plan of a program in the selected college, with details including the program name, planned number of admissions, program code, tuition per 4 years, cutoff exam score and cutoff ranking position last year, and the probability of admission to the program. (Note: the data used in this image is only for demonstration purposes. Screenshots were taken in August 2024.)
• Figure 2: The college application process and timeline in mainland China, as well as the usage of Quark GaoKao app throughout the process.
• Figure 3: The live streaming feature offered by Quark is widely used by our participants, particularly by parents seeking knowledge before the GaoKao exam. Users may check the upcoming live-streaming schedule and select the topics they are interested in. They may also rewatch a past livestream. Users may select a live stream from the schedule to join. In this example, the stream was about strategies to order the list of colleges and programs in the application form to avoid the application being rejected or skipped (if all colleges on the form were full when the application was being processed, the form would be skipped or rejected, leading to the students having no colleges to attend). Users may interact with the streamer through chat messages to ask questions during the live stream. In this example, viewers were asking for suggestions in chat by providing their PA and GaoKao scores.
• Figure 4: Summary of Findings on Parent-Student Dynamics. Internal and external factors influence the design of technology that aims to reduce tension and foster synergy between parents and students.
• Figure 5: Demographic Information of Family Groups. PF1-PF7 represents family members in family-children pairs, PK1-PK7 represents children in family-children pairs. K1-K6 represents individual children, F11-F14 represents individual family members.