Can an AI system reject your CV before a person ever sees it?

A quick rejection email does not show whether a recruiter read your CV. The application may have met a fixed rule, gone through automated screening, or reached a person first.

A job portal can screen an application before a recruiter reads the CV. A rejection sent soon after submission might not reveal what caused it: an AI tool, a pre-set eligibility rule, or a recruiter using screening software.

The Department of Employment and Labour lists sourcing, screening, interview scheduling and applicant-data analysis among recruitment uses of AI. Many applications could also check non-negotiable requirements, such as a qualification, work authorisation, licence or location.

One answer could rule an application out before the CV reaches a recruiter. An early rejection, however, might not prove that AI made the decision. The pressure adds to job-search friction, where an online application can place more work between an applicant and an interview.

Software can screen applications before recruitment staff see a CV. A rejection sent shortly after submission does not prove that AI made the decision, because a fixed rule or a person using software can produce the same email.

Screening software and AI are not the same thing

Employers can use applicant tracking systems, candidate-sourcing tools, video-interview platforms and pre-assessment tools among other technologies. Some employers might also use several platforms that collect and organise applications and leave the decision to a real person.

Meanwhile, another process could apply automatic filters before a shortlist reaches a recruiter, and these details might not be visible on a standard rejection email.

AI already appears in ordinary workplace software, from customer-service tools to meeting records and administration.

What POPIA does and does not cover

Section 71 of POPIA is not a ban on software in recruitment. The provision applies when a decision has legal consequences or substantially affects a person, rests solely on automated processing, and is intended to profile that person.

The Act provides additional safeguards where a contract-related decision relies on appropriate measures to protect the applicant’s interests.

Those safeguards must allow the person to make representations and require enough information about the underlying logic to make those representations.

A human review is crucial because the provision refers to a decision made solely through automated processing.

Recruitment is one part of a wider discussion about AI, personal data and responsibility for decisions that affect people.

AI in a hiring process does not show where human judgment enters. The important question is whether a person reviewed the application before the decision was made.

An application may meet every stated requirement and still be screened out before anyone has weighed the work behind it. The issue might not be whether software has a place in recruitment, but whether an applicant could tell where a rule ends and human judgment begins.