Is AI Biased?

There is controversy that companies using AI recruiting tools are also amplifying existing biases in demographic data. In the United States, demographic questions (such as race, ethnicity, gender, veteran status, or disability status) are typically collected for Equal Employment Opportunity (EEO) reporting and affirmative action compliance; that information is not to be used to determine whether you should be hired.

Ideally:

• The hiring manager does not see these responses.

• Recruiters evaluating candidates do not have access to them.

• The information is stored separately and used only for compliance reporting and audits.

However, applicants are understandably skeptical because many companies now use AI and automated screening tools.

“I can confirm in my experience recruiting and screening applications, that I do not see any demographic data. With your initial application, I do not see or know your age, know your sex, where you were born, your race, or if you were part of any armed forces.”

- Raya Laephuang, Ed.D.

The goal of AI in recruiting should be to reduce administrative burden, not replace human judgment. AI can identify patterns and qualifications, but people are still needed to recognize transferable experience, assess potential, and ensure that qualified candidates aren't overlooked because they don't fit a historical hiring pattern.

“I cannot speak for others, but when I review an application, I am specifically looking at your work history, and to see if you have the basic knowledge and skills that align with the job I am looking to fill. Other than those qualifications, I am also looking for possible certifications you may possess. And if those things meet the criteria, then you’ll be considered and contacted for an interview.”

-Raya Laephuang, Ed.D.

Common Concerns

Can AI see my demographic information?

A well-designed hiring system should not use protected demographic information to rank or reject candidates. Reputable employers generally configure their systems to separate EEO data from the selection process, although that topic is also up for debate.

Can AI infer demographic characteristics anyway?

This is a legitimate concern. Even if race or gender isn't explicitly provided, an AI model could potentially infer characteristics from information such as:

• First name

• Graduation dates (which may indicate age)

• Employment gaps

• ZIP code or location

• Memberships in affinity organizations

• Colleges attended

Researchers and regulators have raised concerns that these indirect signals can introduce bias if AI systems are not carefully tested and monitored.

Should I answer demographic questions?

In most cases, these questions are optional. Choosing "Prefer not to answer" generally should not disqualify you from consideration. Yet, even though AI has been used for years in prescreening and recruiting, we are still in the early stages of navigating this type of terrain.

What Applicants Can Do:

While applicants can't control how an employer configures its AI, they can:

• Emphasize qualifications and measurable accomplishments.

• Keep resumes focused on job-related experience.

• Network with people so there's a greater chance of human review.

• Ask recruiters about their hiring process if they have concerns about AI use.

An HR Perspective:

This is where HR and People Operations play an essential role. Ethical use of AI isn't just about efficiency; it's about governance.

Organizations should:

• Regularly audit AI hiring tools for disparate impact.

• Ensure EEO data is isolated from candidate evaluation.

• Provide meaningful human oversight for hiring decisions.

• Be transparent with candidates about when AI is used.

• Validate that AI recommendations are based on job-related qualifications rather than characteristics that could disadvantage protected groups.

This is becoming a major topic in HR because the question is no longer whether AI should be used; it's how to use it responsibly. The organizations that balance automation with fairness and accountability are likely to earn greater trust from both candidates and employees. The bottom line is AI isn’t the problem; Unchecked AI is the problem.

Most job seekers understand why companies use AI to manage thousands of applications and improve efficiency. The concern isn't that AI exists. The concern is whether applicants have confidence that the technology is being used fairly. If candidates don't understand what information AI evaluates, whether demographic data is separated from hiring decisions, or whether a human reviews the final recommendations, trust begins to erode. Responsible hiring isn't just about implementing AI. It's about auditing it, validating it, and ensuring that human judgment remains part of the process. Technology should help us identify talent, not unintentionally filter it out.

_______

Here are practical steps applicants can take to improve their chances while reducing the risk of being screened out unfairly:

1. Tailor your résumé to each job. Use the same terminology found in the job description when it accurately reflects your experience. AI systems often match skills, certifications, and keywords.

2. Focus on measurable accomplishments. Instead of listing responsibilities, quantify results (e.g., "Reduced onboarding time by 30%" or "Led 25 employee relations investigations annually").

3. Use a simple, ATS-friendly format. Avoid tables, text boxes, graphics, headers/footers, and unusual fonts. Many AI and applicant tracking systems parse plain formatting more accurately.

4. Include relevant skills and certifications. Don't assume the AI will infer your abilities. Clearly list HRIS platforms, software, certifications, and technical skills.

5. Don't rely solely on AI. Whenever possible, network with recruiters, hiring managers, or employees. A referral can help ensure your application receives human attention.

6. Apply through the company's official careers page. Third-party job boards sometimes strip formatting or omit information during the application process.

7. Review your online presence. Ensure your résumé, LinkedIn profile, and portfolio consistently reflect your skills and experience.

8. Highlight transferable skills. If you're changing industries, explicitly connect your previous experience to the new role rather than expecting AI to make that connection.

9. Proofread carefully. Misspellings and inconsistent job titles can reduce keyword matching and make it harder for automated systems to identify your qualifications.

10. Know your rights. In many jurisdictions, employers remain responsible for ensuring AI-assisted hiring complies with anti-discrimination laws. If you believe you've been unfairly screened based on a protected characteristic, you may have legal avenues to raise concerns.