Data issues: Unclear formats, missing info, or inconsistent language confuse the system.

Keyword bias: Overreliance on exact wording; misses context or transferable skills.

Training bias: Past hiring data may encode gender, race, or age bias.

Model limits: AI can misread CVs, infer wrong skills, or ignore new job trends.

Lack of transparency: Decisions may be hard to explain or justify.

No human oversight: Errors or unfair rejections go unchecked.

Technical errors: Parsing, file compatibility, or data quality problems.

Ethical/legal risks: Violating fairness, privacy, or employment laws.

Unfair Exclusion:
The AI system could only read .docx files and ignored all .pdf files. This means that many qualified applicants who submitted their CVs in PDF format were automatically rejected without being considered.

Lack of Equal Opportunity:
Applicants were not treated equally, since only some file types were accepted by the system. Fair recruitment requires giving every candidate an equal chance, regardless of how their CV is formatted.

Bias and Inaccuracy:
Because the AI was trained on old CVs, it might also be biased or outdated in understanding what makes a good candidate today.

Ethical Concerns:
Relying on a system that excludes candidates due to technical limitations is unethical and can damage the company’s reputation.

Unqualified candidates might be hired:
Since the AI ignored all PDF CVs, it only evaluated a small portion of applicants. This means the selected candidates might not actually meet the job requirements.

Qualified applicants would be unfairly rejected:
Many skilled and suitable candidates who submitted their CVs in PDF format would be excluded, reducing the company’s chances of finding the best person for the job.

Bias and discrimination issues:
The recruitment process would become biased, as it unfairly favors applicants who used one file type. This could even be considered discriminatory or unethical.

Damage to the company’s reputation:
If people find out that the company’s hiring process is unfair or faulty, it could harm the company’s public image and trust among applicants.

Wasted time and resources:
Interviewing and possibly hiring unsuitable candidates would cost the company time, effort, and money, and could hurt work quality or productivity later on.

The AI was trained on old CVs, which may not reflect current skills, job titles, or industry trends.

If the old data included mostly CVs from certain backgrounds (e.g., one gender, culture, or school type), the AI may unintentionally favor similar candidates and ignore diversity.

Read the job description carefully. Identify key skills, certifications, job title variants, and include them in your resume — naturally and truthfully. Indeed

Use simple formatting: one column layout, clear headings, common fonts (Arial, Calibri, Times New Roman). Avoid images and extra decorations. SUNY Geneseo+1

Use standard section headings: “Professional Summary”, “Work Experience”, “Education”, “Skills”. This helps ATS recognise the parts. Indeed

Use both acronym & full-form for key certifications/skills (e.g., “PMP (Project Management Professional)”) to cover variations. LHH

Make sure your file is in a compatible format (check the job posting; often .docx or PDF is acceptable, but PDF may sometimes cause issues). Career Services+1

Avoid stuffing irrelevant keywords. It can backfire. The goal is to show real evidence of your skills. LHH

Keep a version of your resume ready for “human review” too — after passing ATS, a human will still look, so clarity and readability matter.