The use of artificial intelligence (AI) in everyday life has become increasingly prominent.
In the workplace, AI holds the potential to fundamentally transform HR practices and workforce management.
The EU Artificial Intelligence Act (the "AI Act")
The AI Act came into force on 1 August 2024 and introduced a risk-based approach to regulating AI systems by using the following levels:
- Unacceptable risk – systems that classify individuals in ways that result in biased outcomes (e.g. discriminatory profiling)
- High risk – systems that make determinations which directly impact individuals (e.g. recruitment and termination of contracts). Most HR functions would be categorised as "high-risk"
- Limited risk – systems developed to perform routine functions (e.g. chatbots)
- Minimal risk – such as spam filters
Provisions concerning AI literacy and the prohibition of certain AI practices came into force on 2 February 2025. On 2 August 2026, the regulations governing high-risk AI systems will apply, marking a critical compliance milestone for employers.
The AI Act has direct implications for businesses operating in the EU, regardless of whether they are providers (those who develop the systems) or deployers (those who use them) of AI systems. The use of AI in the workplace is becoming increasingly common across a range of HR and workforce management functions, including:
- Recruitment
- Assessment of applicants
- Task allocation
- Monitoring employee performance
- Management decisions regarding promotions/terminations
Most employers will fall under the category of deployers, who face fewer obligations than providers. Nonetheless, deployers are still responsible for key duties, including:
- Operating the AI system in accordance with the provider's instructions
- Appointing qualified personnel to manage the AI system
- Ensuring the quality and relevance of input data
- Monitoring the system's performance and reporting any incidents
- Keeping system logs
Recruitment
There has been a significant rise in the use of AI systems throughout various stages of the recruitment process, including:
- Creating job advertisements
- Screening candidates
- Preliminary interviews
- Scheduling interviews
While these tools can offer considerable efficiencies in recruitment, they also carry potential legal risks. From an equality standpoint, employers must be able to justify decisions, such as rejecting a job application, with clear, explainable reasoning. However, this can be difficult when using AI systems, as their internal logic is often opaque, making it challenging to understand how decisions are made.
AI also carries a significant risk of disadvantaging individuals who share certain characteristics protected under section 6 of the Employment Equality Act 1998 (as amended) (1998 Act), such as age, gender, or disability – potentially leading to discrimination.
In general, direct discrimination cannot be justified under the 1998 Act. However, in cases of indirect discrimination, employers may rely on an objective justification defence, provided they can demonstrate that the use of an AI system was a proportionate means of achieving a legitimate aim.
Performance management
AI is also increasingly being used to:
- Assess employee activities and set performance benchmarks
- Identify performance gaps
- Automate performance reviews
While AI offers valuable, data-driven insights, excessive dependence on automated systems for managing performance can undermine trust and confidence between employers and employees.
Risks of using AI in the workplace
Employers should be aware that unintended bias in AI algorithms can arise when these systems are used in areas such as recruitment, performance management, promotions or terminations. If the data used to train an AI model contains bias, the system may replicate that bias, potentially leading to discriminatory outcomes against individuals with protected characteristics. This could result in breaches of the 1998 Act and expose organisations to legal and reputational risk.
In Ireland, compensation for discrimination can be significant: up to two years' remuneration for employees, or up to €13,000 for non-employees (such as job applicants asserting breaches in respect of access to employment).
Under Article 99 of the AI Act, violations can result in severe penalties, including fines of up to €15 million or 3% of global annual turnover. For breaches involving prohibited AI practices, penalties may rise to €35 million or 7% of global turnover, whichever is higher. Employers may need to implement governance structures to ensure that AI technologies are used lawfully and ethically.
Under Article 25 of the AI Act, significant changes to the intended use of any AI system, such as deploying a general-purpose tool (e.g. Chat GPT) in a high-risk context like recruitment, could mean legally becoming its provider. This results in assuming full regulatory responsibility for the system.
Key takeaways for employers
To avoid infringing on employment rights and to ensure compliance, it is essential to establish robust governance and oversight mechanisms. Key steps include:
- Embedding human oversight in all AI system developments through regular auditing, supervision and testing to identify any risk of algorithmic bias
- Providing comprehensive training for staff who interact with AI tools
- Ensuring that input data is accurate, sufficient and in compliance with data protection laws
- Continuously monitoring AI performance and maintaining activity logs
- Informing and consulting with workers on decisions to use AI systems, under Article 26(7) of the AI Act.
- Developing an ethical AI policy which establishes clear guidelines governing the responsible and ethical use of AI systems.
The AI Act marks a significant development in establishing a comprehensive regulatory framework for AI across the EU. As its implementation progresses, employers must remain informed and ready to adjust their practices to meet the new legal requirements.
Contributed by Tara McCormack
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.
