Analysis: The delay gives breathing room for businesses, but the broader imbalance over how monitoring technology is used will not be resolved by the act alone <\/p>\n
Across a growing number of Irish workplaces, AI tools now track what employees do while at work<\/a>. Keystroke counts, mouse movements, login and logout times, how long someone spends idle or away from their screen, and, in some roles, the content and tone of calls and messages can all be tracked. Employers choose the software, decide what it tracks, and configure the thresholds that determine whether someone is flagged as underperforming, at risk of leaving or disengaged. Workers are rarely told how the software reaches those judgments, what data it was trained on, or whether it was tested on people in comparable roles before it was deployed. There is no requirement in Irish law for an employer to disclose how an AI monitoring system was trained, tested or configured.<\/p>\n Irish workers are not without protections in general terms. GDPR gives employees the right to request the data held about them, and unionised workplaces can negotiate how monitoring is used. What no existing Irish law or collective agreement currently requires is that an employer deploying an AI monitoring tool can explain how it was trained, whether it was tested for bias, or on what basis it scores one worker differently from another. Under Article 6 of the EU AI Act<\/a>, AI systems intended to monitor or evaluate the performance and behaviour of workers are generally classified as “high-risk”.<\/p>\n That classification places obligations on both the company providing the system and the employer using it. Workers would have to be told the tool is being used on them before it is switched on, and the employer would have to ensure that a person, not just the software, can review and override its outputs. These are the specific protections that Irish law does not currently provide. Individual workers are rarely in a position to challenge this, because an employee who objects to how their employer’s monitoring software works cannot realistically refuse to be subject to it without putting their job at risk.<\/p>\n We need your consent to load this rte-player contentWe use rte-player to manage extra content that can set cookies on your device and collect data about your activity. Please review their details and accept them to load the content.Manage Preferences<\/a><\/p>\n From RT\u00c9 Radio 1’s The Late Debate, how can we regulate AI? <\/p>\n Those protections were due to apply from 2nd August 2026<\/a>, but in May, EU lawmakers agreed to push the deadline to 2nd December 2027, a sixteen-month delay introduced as part of a broader package of amendments the Commission calls the Digital Omnibus, which adjusted timelines across several parts of the Act. Legal experts<\/a> have described the delay as ‘breathing room’ for business. The European Commission argued that the extra time was needed to give companies the support and technical standards required to comply with the rules.<\/p>\n The European Parliament also pointed to the fact that the technical standards needed to assess compliance were not yet finished. The European Trade Union Confederation<\/a> objected to a separate change in the same package that weakened the Act’s AI literacy requirement, replacing the requirement to ensure a sufficient level of AI literacy among staff with a duty to support its development. The reasons given for the delay focus on giving regulators and businesses more time to prepare, but workers who are screened out of jobs or assessed negatively by monitoring tools in the meantime will have to live with the consequences of that delay.<\/p>\n The technology those protections are intended to regulate is already spreading. Eight of the ten largest private employers in the United States already track the productivity metrics<\/a> of individual workers, many in real time. A study<\/a> of more than 23,000 workers found no evidence that monitoring improves performance, but instead concluded that it increases stress and worsens how workers feel about their jobs. Nonetheless, employers continue to use these tools, as they provide greater visibility and control over how workers spend their time.<\/p>\n The broader imbalance between employers and workers over how monitoring technology is used will not be resolved by the AI Act alone.<\/p>\n The technology is also moving beyond systems that employers choose to install. Smart glasses are a recent example. Unlike monitoring software installed on a company network, glasses can be worn into a workplace by a colleague, a manager or a customer and can record continuously, with some models capable of facial recognition. Few workplaces have a clear consent framework for their use, while retail staff have described being filmed at close range by customers wearing them, with no way to make it stop.<\/p>\n The AI Act’s employment provisions are aimed at specific uses of AI in recruitment, worker management and employment decisions, including monitoring and evaluating workers’ performance and behaviour. That creates a more difficult question around devices such as smart glasses, where surveillance may be carried out by someone other than the employer and may not involve an AI system being used for one of those purposes. Even when the high-risk obligations begin to apply in December 2027, they may not address every form of surveillance that workers encounter.<\/p>\n The obligations coming into force in 2027 will not apply retrospectively, so they will not provide redress for a rejected application, correct a performance score, or explain a decision that has already been made. The delay does, however, leave Ireland with a period in which it can decide how those protections will be put into practice. In September 2025, Ireland became one of the first six EU member states to designate the national authorities responsible for enforcing the AI Act<\/a>.<\/p>\n We need your consent to load this rte-player contentWe use rte-player to manage extra content that can set cookies on your device and collect data about your activity. Please review their details and accept them to load the content.Manage Preferences<\/a><\/p>\n From RT\u00c9 Radio 1’s The Business, the charms of AI<\/p>\n Ireland could use the time between now and December 2027 to require that any hiring or monitoring tool used by an employer in this jurisdiction can show how it was trained, what it measures, and whether the workers it will be used on were informed before it went live. The broader imbalance between employers and workers over how monitoring technology is used will not be resolved by the AI Act alone. Until then, workers will continue to be assessed by systems whose workings they may not understand and whose decisions they may have little opportunity to question.<\/p>\n Follow RT\u00c9 Brainstorm on WhatsApp<\/a> and Instagram<\/a> for more stories and updates<\/p>\n The views expressed here are those of the author and do not represent or reflect the views of RT\u00c9<\/p>\n