From Scores to Careers: Understanding AI’s Role in Supporting Collaborative Family Decision-Making in Chinese College Applications

Research Background and Issues

• Identified Problems and Challenges: The study reveals that in the Chinese college entrance examination (Gaokao) application process, most parents dominate the use of AI tools (e.g., Quark GaoKao), while students typically only participate during the option elimination stage. These AI tools primarily recommend schools and majors based on scores, rarely considering students' long-term career goals. Additionally, issues such as misleading AI recommendations and irresponsible use by third-party consulting agencies persist.
• Importance of the Issue: The Gaokao is an extremely critical societal event in China, decisively influencing students' future career and educational paths. However, current AI tools fail to adequately support collaborative decision-making within families and students' long-term career development, potentially exacerbating social inequities.
• Research Motivation and Related Work: Although there has been research on the role of technology in career development support, studies on AI applications in the score-driven Gaokao context remain limited. The authors aim to explore how AI can support collaborative decision-making in Gaokao applications, providing insights into family-centered design and educational equity.

Solution

• Proposed Methods or Solutions: The study conducted in-depth interviews with 32 participants (including students, parents, and experts) to analyze the role of Quark GaoKao in multi-party decision-making and proposed design recommendations to improve these tools. Specific methods include identifying usage patterns of AI tools, analyzing participant feedback, and exploring the dynamics of family collaborative decision-making.
• Innovations:
  • Introducing the concept of family-centered design applied to AI technology, encouraging designs that emphasize information symmetry, emotional support, and generational differences between parents and students.
  • Combining the score-driven Gaokao context with analyses of family dynamics and career development considerations, a perspective rarely addressed in current research.
• Implementation Steps and Techniques:
  • Data Collection: Conducted semi-structured interviews via WeChat with 32 participants representing different family roles.
  • Data Analysis: Used qualitative thematic analysis to code interview transcripts, identifying participants' usage, trust, and challenges regarding AI tools in the Gaokao application process.

Research Findings

• Specific Findings:
  1. Parent Dominance and Limited Student Participation: Parents often handle data verification and information triangulation tasks, while students mainly eliminate options they are uninterested in.
  2. Limitations of AI Tools: Current AI tools focus primarily on score optimization, neglecting students' personalized interests, long-term career development, generational differences within families, and issues of information asymmetry.
  3. Impact of Resource Disparities: Low-resource families exhibit weaker critical thinking and usage capabilities regarding AI recommendations, with limited technical literacy and social capital exacerbating educational inequities.
  4. Misleading Information and Diffused Responsibility: Some institutions and live-streamed content use AI as a gimmick to attract parents but often provide shallow and homogenized information.
• Comparison with Existing Solutions and Advantages:
  • Proposed improvements to tool design, such as adding career development-related features, personalized recommendations, and "scenario exploration" options to enhance the rationality and transparency of AI tools in decision-making.
  • Emphasized the potential of AI tools to foster parent-child dialogue and emotional interaction within families, rather than solely optimizing scores.
• Experimental or Evaluation Results:
  • The study uncovered potential tensions between parents and students due to information and goal asymmetry, while expert data validation further confirmed the shortcomings of current AI recommendation algorithms.
  • Some resource-rich families successfully combined books, live-streams, and expert advice to validate AI results, though this process was time-consuming and complex.
• Limitations and Future Directions:
  • The study sample is concentrated in northern China; future research should expand to southern regions or other policy contexts.
  • Focus on students (particularly those under 18) to improve AI tools' engagement and accuracy in education and career planning.
  • Exploring ways to better support low-resource families through technology to address educational equity will be a crucial direction for future research.

Through the above analysis, this study provides a comprehensive examination of the use of AI tools in the context of the Gaokao and offers valuable design recommendations for optimizing these tools, contributing to the integration of technology and educational equity. These insights can also be applied to the design and practice of other high-stakes educational decision-making environments.