AI Hiring Tools and Discrimination: Employment Equality Acts Points for Irish Employers

If an AI tool helps you screen, rank or select job applicants, the Employment Equality Acts still apply to the hiring process. The Acts do not carve out decisions shaped by software, and the EU AI Act separately treats many recruitment tools as high-risk. This guide explains the equality points an Irish employer should have in mind when using AI in hiring: which grounds are protected, what direct and indirect discrimination mean as the official pages describe them, and what the AI Act says about data quality and human oversight. It is general information, not legal advice, and we do not cite studies or statistics on how often bias occurs. For the AI Act side of recruitment tools, see our guide to AI in recruitment and HR under the EU AI Act.

What the Employment Equality Acts cover

Citizens Information (page last edited 31 July 2023) explains that the Employment Equality Acts 1998 to 2015 cover areas including equal pay, recruitment, training and work experience, promotion or re-grading, terms and conditions, classification of posts, collective agreements and dismissal. It lists nine protected grounds: gender, civil status, family status, sexual orientation, religion, age, disability, race, and membership of the Traveller community. It notes that age protection does not cover anyone under 16, that pregnancy-related discrimination is treated as gender discrimination, and that harassment and victimisation are also covered.

The Workplace Relations Commission (WRC) page on employment equality says the Acts cover employees in the public and private sectors as well as applicants for employment and training. It says the Acts outlaw discrimination in work-related areas such as pay, vocational training and access to employment, and that discriminatory advertisements are outlawed. The page shows no last-updated date. Note that the Irish Human Rights and Equality Commission (IHREC) page we read lists a tenth item, housing assistance payment, alongside the nine grounds, and is not consistent in its own count, so check the current position on the grounds relevant to your hiring.

The key point for AI is simple. The Acts apply to recruitment and to applicants. An automated tool that filters out applicants is part of the recruitment process, and the official pages we read do not suggest otherwise. They also do not say how the Acts apply to a tool bought from a vendor, so we cannot say more on that from these sources.

Direct and indirect discrimination in an AI context

Citizens Information describes direct discrimination as requiring a direct comparison, for example between a person with a disability and a person without one, and indirect discrimination as occurring when a policy that seems neutral still disadvantages a particular group. IHREC's page similarly describes indirect discrimination as an apparently neutral provision, criterion or practice that puts a person with a protected characteristic at a particular disadvantage compared with others, and says it is not discrimination if it can be justified. That page cuts off before listing the justification criteria, so we do not set them out.

Our suggestion for using these descriptions: treat an AI hiring tool as a set of criteria and practices. A scoring rule that looks neutral, such as a preference for uninterrupted work history or particular keywords, is still a criterion applied to every applicant. If it disadvantages people with a protected characteristic, the indirect discrimination question arises, however sophisticated the tool. That is our reading of how the descriptions apply, not a statement from the sources.

Disability and reasonable accommodation

Citizens Information says employers must make reasonable accommodation for employees with disabilities, including in access to employment and participation in promotion and training. For hiring, our suggestion is to check whether any AI-enabled step, such as a timed online assessment, a video interview scored by software or a chatbot application form, can be completed by a candidate who needs an adjustment. Make sure candidates know how to ask for one and that a person handles the request. Our guide to reasonable accommodation under the Employment Equality Acts covers the duty itself in more depth.

What the AI Act says about recruitment tools and data quality

The European Commission's AI Act Service Desk reproduces the Regulation (EU) 2024/1689 text, and says its page is based on the EUR-Lex consolidated version as at 27 July 2026, with Digital Omnibus on AI changes marked. Annex III point 4(a) lists AI systems intended for recruitment or selection of natural persons, in particular to place targeted job advertisements, to analyse and filter job applications and to evaluate candidates. Point 4(b) lists systems used to make decisions affecting terms of work-related relationships, promotion or termination, to allocate tasks based on individual behaviour or personal traits, or to monitor and evaluate performance and behaviour.

Article 10, as shown on the Service Desk, sets data governance requirements for the training, validation and testing datasets of high-risk AI systems. These include examining datasets for biases that could affect health and safety, fundamental rights or lead to prohibited discrimination, and taking measures to detect, prevent and mitigate those biases. Datasets must be relevant, sufficiently representative and, as far as possible, free of errors and complete for the intended purpose. The Service Desk page marks paragraph 5 as deleted and lists a new Article 4a on processing special categories of personal data for bias detection and correction; we have not read Article 4a, so we say nothing further on it.

The key practical point is who does what. Article 10 is framed as a requirement for the high-risk system itself, and the Service Desk page for Article 16 lists, among providers' obligations, ensuring their systems meet the Section 2 requirements; in our reading, an employer buying a tool is usually a deployer rather than the party that builds the datasets. The Service Desk summary of Article 26 sets out deployer duties: using the system according to the provider's instructions, assigning human oversight to people with the necessary competence, training and authority, and, where the deployer controls input data, making sure it is relevant and sufficiently representative for the intended purpose. It also says deployers using a high-risk system at the workplace must inform workers' representatives and affected workers beforehand, and that deployers of Annex III systems that make or help make decisions about natural persons must inform those people they are subject to the system. For more on those duties, see our guide to deployer obligations and human oversight under Article 26.

On timing, the Commission's Navigating the AI Act page (updated 7 August 2026) says high-risk rules are now slated to apply from 2 December 2027, linked to the Digital Omnibus. Treat that as the position on that page on that date and check the current position. Meanwhile, the Employment Equality Acts already apply to your hiring.

A practical checklist for AI in hiring

A practical approach, labelled as our suggestion and not as a legal requirement:

• Know the tool. Record what each AI step does, what data it uses and what the vendor says about its intended purpose.
• Ask the vendor. Request documentation on how the system was tested for bias and on what data, and keep the answer on file.
• Keep a person in the decision. Make sure the reviewer can see why an applicant was ranked and has the time and authority to override.
• Check outcomes. Periodically review who is progressing through each stage. Where patterns look odd for any protected ground, investigate and pause use if needed.
• Offer adjustments. Build a clear route for candidates to request one.
• Tell people. Be open with applicants and workers about where AI is used.
• Keep records. You may need to explain your process if a complaint is made.

If a complaint is made

Citizens Information (page last edited 10 March 2025) says discrimination complaints in employment go to the WRC through its online complaint form, and may be resolved by mediation or adjudication, with appeals to the Labour Court. Employment complaints must be brought within 6 months of the discrimination, extendable by a further 6 months for reasonable cause. Redress can include compensation, an order for equal treatment and an order that someone take a specified action. The page does not explain who must prove what, so we do not describe the burden of proof here. Check the WRC's own guidance if a claim arises.

Learnsignal's offices are off Baggot Street in Dublin, and we offer in-person training from there. For current details, see our CPD pages.

Frequently asked questions

Do the Employment Equality Acts apply if an algorithm rejects a candidate?

The Acts cover recruitment and applicants. The pages we read do not distinguish between human and automated steps in the recruitment process, and we suggest treating both the same way.

Is an AI hiring tool high-risk under the AI Act?

Annex III point 4 lists systems intended for recruitment or selection, including analysing and filtering applications and evaluating candidates. The Article 6 page also says an Annex III system that performs profiling of natural persons is always considered high-risk. Check the current text, as Digital Omnibus changes are marked on the Commission's page.

Who is responsible for checking the tool for bias, the vendor or the employer?

Article 10 sets data governance requirements, including bias examination, for high-risk systems; the Service Desk page does not say who carries them, so ask the vendor. Deployers have their own duties, including human oversight and input data quality where they control it. The Equality Acts pages we read do not address how liability is shared between employer and vendor.

What should we do if we spot a pattern that looks unfair?

Our suggestion is to pause reliance on the affected step, involve a person to review affected candidates, ask the vendor for an explanation and take legal advice.

AI can speed up screening, but the duty not to discriminate stays with the employer. Understanding your tool, keeping a person in the loop and recording what you do puts you in a better position if a question is raised.