XIAOXI XU | CECILIA

 

WORK

RESEARCH






APPLAI| AI求职助手

A Transparent Hiring Platform that Holds Companies Accountable and Exposes Bad Apples
2025.9-12 HCI Studio

A fair and transparent hiring assistant that helps applicants decode AI-driven evaluations and uncover bad apples through clearer signals, documentation, and guidance.

Job seekers often face opaque AI hiring systems. AppIAI brings clarity by revealing how algorithms evaluate applications and by guiding users toward trustworthy opportunities.

Product Pitch

12/05/2025Cornell University CIS Ithaca,NY




User Research


Literature Review We examined three recurring challenges in AI-mediated hiring: reliance on voluntary employer compliance, limited transparency for applicants, and persistent enforcement gaps in regulatory oversight. We discovered that existing approaches largely prioritize employer needs, leaving applicants with limited visibility into hiring decisions, little avenues for recourse, and minimal voice in the process.



Stakeholder Interview We conducted nine semi-structured interviews with job seekers (n=2), hiring professionals (n=4), and industry experts (n=3). All participants provided informed consent prior to participation, and identities are anonymized in reporting. Data collection used synchronous video interviews (45-60 minutes) and asynchronous written questionnaires with follow-up clarifications. Sessions through detailed note-taking.



Low-Fi prototype We conducted a prototype evaluation with 3 job-seeking students actively applying for internships or full-time roles, along with 3 studio classmates who provided peer critique. Participants completed task-based walkthroughs using a paper prototype and a think-aloud protocol.


Expert Consultation We consulted with two subject matter experts:
  • An AI law professor specializing in automated decision-making and employment regulation
  • An ILR professor with expertise in hiring practices and labor dynamics


User
Testing We conducted a 45-minute moderated usability test using a high-fidelity Figma prototype. Participants completed scenario-based tasks aligned with key stages of the hiring journey, including document review, application tracking, company evaluation, and access to support features. Sessions followed a think-aloud protocol, with one moderator and one observer. The notes were later clustered and analyzed through affinity diagram.

Participants (n=6) were all active job seekers, including five international students and one U.S.-based student, selected to reflect users most affected by AI-mediated hiring. All participants were first-time users and received minimal onboarding.



AI
UI Audit To evaluate the usability, accessibility, and visual consistency, we conducted a structured UI audit supported by ChatGPT as an expert-review aid. The system was prompted to evaluate the main user flows using established frameworks, including Nielsen’s usability heuristics, Material Design 3 guidelines, WCAG 2.2 accessibility standards, and consistency with the ApplAI brand system.

The UI audit shows the design is conceptually strong but needs targeted refinements.



1. Introduction
This project explores how accountability in AI-mediated hiring can be established without relying on voluntary employer compliance. Through stakeholder interviews with job seekers, hiring professionals, and industry experts, alongside literature review, iterative prototyping, and evaluation, we developed ApplAI, a platform designed to rebalance power toward applicants.

Unlike existing tools that help job seekers optimize resumes to pass algorithmic filters, ApplAI emphasizes visibility into company hiring practices and provides access to legal guidance when outcomes appear unfair. This approach supports applicants in understanding hiring decisions and recognizing when formal recourse may be appropriate.




2. Analysis & Definition
We developed and evaluated several design concepts based on our objectives and user persona. Based on alignment with user needs and design objectives, we selected a platform-based concept that integrates hiring insights, application tracking, and support for responding to unfair outcomes.



We then refined our approach through low-fidelity prototype testing with job seekers and feedback from subject matter experts, resulting in the high-fidelity design of ApplAI.




3. Design & Implementation ApplAI is a fair and transparent hiring assistant that helps applicants decode AI-driven evaluations and uncover bad apples through clearer signals, documentation, and guidance.

Job seekers often face opaque AI hiring systems. AppIAI brings clarity by revealing how algorithms evaluate applications and by guiding users toward trustworthy opportunities.




4. Discussion
 This project demonstrates how applicant-facing design can support AI-based hiring by making automated evaluation clearer and easier to understand from the applicant’s point of view. Instead of expecting transparency only from employers or regulators, ApplAI looks at ways to support applicant agency even when information is limited or unevenly shared.

This work shows that designing for accountability is less about revealing every system detail and more about helping people understand how decisions are made. It explains where automation is used, records outcomes, and helps applicants judge employer practices. It also highlights that transparency should be presented carefully and with empathy, so it supports applicants without adding extra stress.

Although ApplAI is only a prototype, it offers a practical way to make hiring more accountable to applicants, alongside rules and company policies. As automated hiring becomes more common, this work suggests that real accountability between people and AI can start with the design of the interface. Good design can help applicants move from being passive subjects to active, informed participants in the hiring process.