Inside Story
AI in the Workplace: Useful Assistant or Growing Employment Risk?
Artificial intelligence is becoming part of everyday working life. Employers are using AI to screen CVs, draft correspondence, summarise meetings, analyse workforce information and assist with employee relations matters. Employees are also using it to prepare grievances, appeals, responses to disciplinary allegations and submissions to the Workplace Relations Commission.
Used responsibly, AI can improve efficiency and help people organise complex information. Used without proper oversight, it can produce inaccurate advice, invented case law, discriminatory outcomes, data protection breaches and workplace decisions that cannot be properly explained or defended.
With new Irish and EU legislation now in place, employers should no longer treat workplace AI as simply an IT matter. It is increasingly an issue of employment law, equality, data protection, governance and fair procedure.
Inside This Edition
1. AI in grievances, complaints & disciplinary procedures
2. The risks for employees using AI-generated submissions
3. AI in recruitment and candidate selection
4. The latest EU AI Act implementation dates
Ireland’s Regulation of Artificial Intelligence Act 2026
5. Case Law Spotlight: Oliveira v Ryanair DAC
Practical actions for employers
Trending Now
AI in Grievances, Complaints and Disciplinary Procedures
AI tools can help HR teams and managers organise timelines, identify gaps in documentation and prepare first drafts of letters or meeting questions. However, they should not be used to determine whether an allegation is proven, assess a person’s credibility or decide what disciplinary sanction should apply.
An AI system does not understand workplace context in the same way as a trained investigator or decision-maker. It may overlook relevant evidence, misunderstand the applicable policy or create conclusions that sound convincing but are not supported by the facts.
This presents particular risks where AI is used to:
-> Generate investigation findings
-> Assess the credibility of witnesses
-> Recommend disciplinary sanctions
-> Summarise lengthy witness statements
-> Draft grievance or disciplinary outcomes
-> Analyse an employee’s tone, behaviour or perceived emotional state
-> Predict whether an employee is likely to raise a complaint or leave the organisation
Under the EU AI Act, certain uses of AI to infer a person’s emotions in the workplace are prohibited, subject to limited exceptions. This prohibition has applied since 2 February 2025.
Employees Are Using AI Too
Employees are increasingly using AI to draft grievances, complaints, appeals and WRC submissions. AI can help an employee structure their concerns or improve the clarity of a document. The danger arises when the employee relies on AI-generated legal arguments, case citations or factual conclusions without checking them.
On 26 May 2026, the Workplace Relations Commission updated its official guidance for parties using AI to prepare written submissions or evidence. The guidance confirms that the WRC will treat the submission as the party’s own work, even where AI assisted with its preparation. The individual remains responsible for ensuring that the content is relevant, current and accurate.
The WRC identifies several specific risks:
-> AI may provide incorrect legal advice.
-> It may invent legislation, decisions, quotations or case references.
-> It may produce arguments that sound authoritative but do not address the actual dispute.
-> Incorrect material can waste hearing time and create additional work for the other party.
-> Information entered into public AI tools may create privacy or confidentiality risks.
The WRC advises users to verify every legal authority and ensure they understand and can explain everything included in their submission. It also suggests that parties consider voluntarily disclosing that AI assisted with the drafting
AI in Recruitment
Recruitment is one of the areas most directly affected by the EU AI Act.
AI systems may be used to:
-> Screen or reject CVs
-> Rank or shortlist applicants
-> Match candidates to vacancies
-> Analyse online interview responses
-> Target job advertisements
-> Score aptitude or personality assessments
-> Recommend hiring decisions
The WRC confirms that many systems used for recruitment, selection and worker management may be treated as high-risk. Examples include systems used to screen applications, rank candidates, allocate work, monitor performance, assess promotion opportunities or support decisions affecting terms of employment.
Employers should be particularly cautious where an AI tool may reproduce biases contained in historical recruitment data. A system trained on previous hiring decisions may unintentionally disadvantage candidates based on gender, age, disability, race, family status or another protected ground.
Using an external recruitment platform does not necessarily transfer all responsibility to the supplier. Under the EU AI Act, an employer using a system may have responsibilities as a deployer, while additional obligations can arise if the employer develops, substantially modifies or rebrands the system.
Recruitment Questions Employers Should Ask
-> Does the system automatically reject or rank candidates?
-> What information is used to make the recommendation?
-> Has the provider tested for discriminatory outcomes?
-> Is a trained person reviewing the result?
-> Can the organisation explain the decision to an applicant?
-> Are applicants informed appropriately about the processing of their information?
-> Is sensitive candidate information being entered into a public AI tool?
Many AI-powered workplace tools use profiling to assess or predict things about individuals, such as their performance, reliability, behaviour, attendance patterns or even their likelihood of leaving an organisation. By analysing large volumes of personal data, AI systems can identify patterns, generate scores and inform decisions relating to recruitment, promotion, performance management, workforce planning and employee monitoring.
While profiling can improve efficiency and consistency, it also carries significant risks. Employees and candidates may not fully understand how they are being assessed or what factors are influencing the outcome. There is also a risk that AI systems may reinforce existing biases or discriminatory patterns if not properly designed and monitored. Profiling is often used as part of automated decision-making, where technology makes decisions with little or no human involvement, making transparency, fairness and human oversight increasingly important in the workplace.
Human involvement must be meaningful. A manager simply approving an AI recommendation without critically reviewing it is unlikely to provide adequate protection against unfair or discriminatory outcomes.
If AI Gets It Wrong: Understanding the Risks for Employers
While the EU AI Act, implemented in Ireland through the Regulation of Artificial Intelligence Act 2026, which was enacted on July 21, 2026, introduces new safeguards around workplace AI. Employers should be aware that the legal risks associated with AI-driven decision-making extend far beyond the legislation itself. Employees in Ireland can already challenge decisions they believe are discriminatory, unfair, inaccurate, or lacking appropriate human oversight through internal grievance procedures, the Workplace Relations Commission (WRC), or the Data Protection Commission (DPC), depending on the nature of the complaint. Where AI influences recruitment, performance management, promotion, disciplinary action, redundancy selection, or dismissal decisions, employers remain fully accountable for the outcome. The EU AI Act strengthens these existing protections by requiring greater transparency, human oversight, and accountability, while increasing the potential regulatory, financial, reputational, and employee relations risks for organisations that fail to govern AI appropriately. Ultimately, AI may assist decision-making, but responsibility for those decisions will continue to rest with the employer.
The AI Act classifies AI systems used in areas such as recruitment, promotion, performance management and termination as high-risk, requiring employers to implement transparency, accountability and human oversight measures. These principles were further strengthened by the CJEU’s February 2025 ruling in CK v Dun & Bradstreet Austria, which confirmed that individuals have a right to a meaningful explanation of AI-assisted or automated decisions that affect them. In practice, employers must be able to explain how an AI-driven outcome was reached and ensure there is a genuine opportunity for human review and challenge.
How EMGHR Can Help
Before using an AI-assisted HR document, ask:
“Could I explain and defend every factual statement, legal reference and conclusion in this document before the Workplace Relations Commission?”
If the answer is no, the document is not ready to be used.
EMGHR can support organisations with:
-> AI workplace policies
-> AI and HR risk assessments
-> Recruitment process reviews
-> Grievance and disciplinary procedures
-> Workplace investigations
-> GDPR and HR compliance reviews
-> Manager training
-> WRC preparation and case-management support
-> HR documentation and policy updates
AI can bring real benefits to the workplace, but speed must not come at the expense of fairness, accuracy or accountability.
For practical guidance on introducing AI responsibly within your organisation, contact EMGHR.
Case Law
Case Number: ADJ-00055225
Fernando Oliveira V Ryanair DAC
In Fernando Oliveira v Ryanair DAC, a cabin crew employee alleged discrimination on the grounds of race and family status, together with victimisation, harassment and sexual harassment under the Employment Equality Acts. The complaint arose from workplace disputes involving another employee, disciplinary investigations conducted by Ryanair and alleged barriers to training and promotion opportunities. The hearings took place on 20 March 2025 and 12 June 2025, with the decision issued on 1 October 2025 by Patricia Owens, Adjudication Officer of the Workplace Relations Commission.
The complainant submitted extensive written arguments and sought compensation of €170,000 together with the removal of disciplinary sanctions and a review of Ryanair’s internal processes. A significant aspect of the case was that the complainant relied on numerous legal authorities and WRC decisions that the respondent argued were either misquoted, irrelevant, incorrectly described or, in some instances, non-existent. The respondent submitted that the documents appeared to have been generated with the assistance of AI and contained what were described as “phantom citations”. The complainant acknowledged he may have used AI and argued that he was entitled to do so.
The Adjudication Officer ultimately rejected all elements of the complaint. She found that the complainant failed to establish a prima facie case of discrimination, failed to provide evidence linking alleged treatment to his race or family status, and failed to provide cogent evidence supporting claims of victimisation, harassment or sexual harassment. The complaint was found to be not well founded
The AI Issue
While the substantive discrimination claim failed, the most interesting aspect of the decision for employers is the commentary relating to AI-generated submissions.
The Adjudication Officer stated that she was not particularly concerned about whether AI had been used in preparing submissions. However, she emphasised that parties remain fully responsible for ensuring that any submissions made to the WRC are accurate, relevant and not misleading. She found that the complainant’s submissions contained citations that were irrelevant, misquoted and, in many cases, did not exist. The decision notes that considerable time was spent by both the respondent and the WRC attempting to verify the authorities relied upon.
The Adjudication Officer concluded that the attempt to rely on “phantom citations” and unsupported allegations could only be described as “egregious” and an “abuse of process”.
This decision has since become one of the key Irish employment law authorities highlighting the dangers of relying on AI-generated legal content without proper verification. It was subsequently referenced within the WRC’s official guidance on the responsible use of AI in WRC proceedings.
Key takeaway for employers:
The decision does not establish that using AI in an employment claim is automatically improper. Instead, it demonstrates that:
-> AI cannot discharge the burden of proof.
-> A polished submission cannot replace supporting evidence.
-> Parties must understand and defend the arguments they submit.
-> Employers benefit from maintaining a clear and itemised documentary record.
-> Evidence, contracts, payslips and contemporaneous correspondence will generally carry greater weight than unsupported AI-generated assertions.
The WRC’s subsequent guidance reinforces this point. AI can support drafting and research, but the person submitting the material remains fully responsible for its accuracy and relevance.
EMGHR Comment:
Artificial Intelligence is becoming a common feature in workplace disputes. Employees are increasingly using AI tools to draft grievances, disciplinary appeals, protected disclosures and WRC complaints. Employers are also using AI to draft investigation reports, correspondence and policy documentation.
This case highlights an important distinction:
Using AI is not the problem. Using AI without verifying the output is.
AI can be an extremely useful administrative and drafting tool, but it cannot replace evidence, legal analysis, critical thinking or professional judgement.
For employers, the key takeaway is not to fear AI but to govern it properly. Organisations should ensure that:
-> AI-generated HR documents are reviewed by a competent person.
-> Legal authorities and references are checked before reliance is placed upon them.
-> Investigations remain evidence-based and independently conducted.
-> HR teams and managers receive training on the appropriate use of AI.
-> Clear AI usage policies are introduced across the organisation.
As the use of AI continues to evolve, employers who combine technology with strong governance, fair procedures and sound HR practices will be best positioned to manage workplace disputes effectively.
Implications for Employers
This case demonstrates several important lessons for employers:
-> AI Does Not Replace Evidence
-> Regardless of how professionally drafted a submission may appear, the WRC will assess the evidence supporting the allegations. Assertions, speculation and unverified AI-generated content will not satisfy the legal burden of proof.
-> Documentation Matters
-> Ryanair successfully defended the complaint because it was able to demonstrate that complaints had been investigated, disciplinary procedures had been followed, appeals had been conducted and administrative errors had been corrected when identified. Employers should maintain robust records of investigations, disciplinary meetings, decisions and appeals.
-> AI-Generated Grievances and Complaints Are Increasing
-> Many HR professionals are now seeing grievances, appeals and WRC submissions drafted using AI tools. While these documents often appear comprehensive, employers should not assume that legal authorities, factual assertions or references relied upon are accurate. Each point should be independently reviewed and verified.
-> Fair Procedures Still Win
-> The decision reaffirmed that employers who follow documented procedures, provide employees with opportunities to respond, conduct investigations and maintain records remain in a strong position when defending complaints.
-> AI Governance Is Becoming an HR Issue
-> As AI becomes more widely used for grievances, disciplinary appeals, investigations and recruitment processes, organisations need clear guidance on how employees and managers can use AI responsibly and appropriately.
Partnership
Introducing EMA: The All – in – One Platform for Growing Businesses
Running a successful business means managing people, clients, compliance, recruitment, sales, and operations, often across multiple disconnected systems.
EMA brings everything together in one powerful platform. Designed specifically for growing organisations, EMA seamlessly connects Customer Relationship Management (CRM), Human Capital Management (HCM), Applicant Tracking (ATS), Learning & Development, and business operations into a single source of truth.
By eliminating duplicate data, streamlining processes, and providing real-time insights, EMA empowers business owners and managers to focus less on administration and more on building high-performing teams, strengthening client relationships, and driving sustainable growth.
Whether you’re attracting top talent, managing employee development, improving sales performance, or enhancing operational efficiency, EMA provides the tools to help your business thrive.