JobBecome如何使用AI

AI帮助求职者和企业提高效率，但最终决定始终由人来做

1. AI assists; people decide

No application is rejected automatically. Every application reaches the employer, and the employer decides.
Candidate–job comparisons are shown as groups (strong / partial / not enough data / weak) with a reason for each requirement — no numeric score is used to judge people.
A requirement with no evidence found shows "no evidence found", not "failed", and we suggest asking the candidate.

The main basis is the requirements the employer confirmed, compared one by one with the profile and CV, each with supporting evidence.
AI only checks items the rules cannot answer, and must quote text from the candidate's history — if that text does not actually exist, the result is not used.
Answers the candidate gives about themselves are labelled "self-reported" and are not counted as meeting a requirement until the employer confirms.
AI can help draft job ads, but the employer must review them first, and the system removes numbers the AI made up, such as years of experience the employer did not specify.

The system does not accept requirements or screening questions about gender, age, religion, marital status, health or appearance.
Text in a candidate's history about these topics is never used or quoted as evidence.

AI runs on JobBecome's own machines; candidate data is not sent to outside AI services.
Names, email addresses and phone numbers are removed before AI reads a history or CV.
Profile details are attached to an application only when the candidate chooses, and profile visibility has 4 levels — contact details are shown only to companies the candidate accepts.
Recommended jobs and the "you may be a strong applicant" badge are calculated from the user's own data when they open the page, are not stored or shared with employers, and always show the reasons.

Requirement comparisons are measured against real recruiters' decisions, with a focus on the system never overrating a candidate.
The data used for measurement is anonymised.