Hiring is a high-volume, high-stakes process. A single job posting can attract hundreds or thousands of applications, and AI promises to help recruiters find the best candidates faster. But hiring decisions shape people’s livelihoods, and algorithmic tools can reproduce or amplify discrimination. That makes HR one of the most regulated areas of AI.

Where AI is used in HR

Sourcing. Tools search professional networks and databases for potential candidates and draft outreach messages.

Job descriptions. AI drafts postings and flags language that may discourage certain groups from applying.

Screening. Applicant tracking systems (ATS) organize applications; AI features rank résumés against job requirements, extract skills and filter by criteria.

Assessments. Online tests, games and structured questionnaires, sometimes scored with machine learning.

Interviews. Scheduling assistants, chatbots that answer candidate questions, and video interview platforms. Some tools have analyzed facial expressions or voice — a practice widely criticized as scientifically unsupported, and which some vendors have dropped.

Employee lifecycle. Onboarding assistants, internal job matching, learning recommendations, attrition prediction and HR helpdesk chatbots.

The bias problem

A well-known example: Reuters reported in 2018 that Amazon had scrapped an experimental recruiting tool after finding it penalized résumés that included the word “women’s” (as in “women’s chess club captain”). The model had learned from a decade of past hiring data dominated by men. The case illustrates a general risk: models trained on historical decisions learn historical biases.

Other risks:

  • proxies such as postcodes, gaps in employment or certain schools that correlate with protected characteristics;
  • disability discrimination when tools penalize candidates who interact differently, for example in timed tests or video analysis;
  • opacity: candidates rarely know how they were evaluated;
  • errors in parsing résumés with unusual formats.

The rules

  • New York City Local Law 144, enforced since July 2023, requires employers using automated employment decision tools for hiring or promotion in the city to commission an annual independent bias audit, publish a summary of the results, and notify candidates.
  • The EU AI Act classifies AI used for recruitment, selection, promotion, termination, task allocation and worker monitoring as high-risk, requiring risk management, data governance, human oversight, transparency and logging. It also prohibits emotion recognition in the workplace, with limited exceptions for medical or safety reasons.
  • Anti-discrimination laws apply regardless of technology: employers remain responsible for discriminatory outcomes, even when caused by a vendor’s tool.
  • Data protection laws like the GDPR give candidates rights regarding automated decisions and their personal data.

Good practices for employers

  1. Use AI to support, not replace, human hiring decisions.
  2. Validate that assessments actually predict job performance.
  3. Test for adverse impact across groups before and during use.
  4. Provide accommodations and alternatives for candidates with disabilities.
  5. Tell candidates when and how AI is used.
  6. Require transparency and audit support from vendors.

Questions candidates can ask

  • Is AI used to screen or assess my application, and how?
  • Will a person review my application before a decision?
  • Can I request an accommodation or an alternative assessment?
  • How long will my data be kept, and can I ask for it to be deleted?

Beyond hiring

AI in HR increasingly extends to managing current employees: scheduling, performance analytics and productivity monitoring. These uses raise concerns about surveillance and stress at work, and in the EU, AI used to monitor and evaluate workers is also high-risk.

What to watch next

  • EU AI Act obligations for employment AI coming into force.
  • More local and state bias-audit laws for hiring tools.
  • Litigation over discriminatory AI screening.
  • Candidates’ use of AI to apply en masse and employers’ responses.

Key takeaways

  • AI is used in sourcing, screening, assessments, interviews and employee management.
  • Amazon scrapped a recruiting tool that learned to penalize résumés mentioning “women’s”.
  • NYC Local Law 144 requires annual independent bias audits of automated hiring tools.
  • The EU AI Act classifies employment AI as high-risk and bans workplace emotion recognition.
  • Employers remain responsible for discriminatory outcomes of vendor tools.

Frequently asked questions

Is my CV being read by AI?

Often software organizes and may rank applications. Many employers still have humans review shortlisted candidates.

Can I ask whether AI was used in my application?

In many jurisdictions you have the right to information about automated processing; some laws require employers to notify you.

Are AI video interviews analyzing my face?

Some tools have done so, but the practice has been widely criticized and is restricted in places like the EU. Ask the employer what is assessed.