Regulators across Europe have begun enforcing fresh transparency mandates under the EU AI Act, forcing HR leaders to confront how artificial intelligence operates inside their workforces right now rather than waiting until 2027.
The enforcement push arrived on 2 August 2026, marking a pivotal shift for employers throughout the UK and beyond. Although the legislation originates in Brussels, its reach extends to any company developing or deploying large language model tools, regardless of location.
Regulators can now crack down on the AI literacy obligation that has technically existed since February 2025. Meanwhile, many requirements tied to high-risk AI systems used in employment decisions have been postponed until December 2027. That delay invites a dangerous assumption: that preparation can wait. Industry observers argue the opposite.
The literacy requirement carries significant weight. It demands far more than showing employees how to operate AI tools. Organisations must ensure workers understand where AI gets deployed, recognise its limitations, know when to challenge outputs, and determine where human judgment should override automated results.
Why does this matter for HR specifically? AI adoption rarely spreads evenly across a workforce. Without a shared foundational understanding, organisations risk inconsistent practices, eroding trust and diminishing psychological safety among employees.
The transparency rules add another critical layer. As AI becomes woven into recruitment, learning platforms, employee services and internal communications, workers deserve clarity about when AI influences their experience and where human oversight still governs decisions. Transparency serves as the bedrock of workforce trust, not merely a compliance checkbox.
A growing tension deserves attention. Generative AI now helps employees draft grievances, appeals and workplace requests that appear authoritative but may cite irrelevant legislation or policies. Managers increasingly lean on AI to craft responses and support decision-making. While productivity gains are real, AI cannot replace contextual understanding, empathy or professional judgment.
HR’s mandate extends beyond discouraging or encouraging AI use. The function must help managers and employees discern where AI genuinely adds value and where decisions demand informed human intervention.
Several questions demand answers long before December 2027. Where does AI already influence recruitment, performance reviews and development pathways? Who bears accountability for AI-informed decisions? Could those decisions withstand scrutiny if challenged? How can organisations ensure AI promotes fairness rather than deepening existing inequalities?
These are people questions, not technical ones. HR sits squarely at the centre of responsible AI adoption. The organisations best positioned for 2027 will not necessarily spend the most on AI technology. They will invest in capability, transparency and sound judgment. Compliance, when done properly, becomes the byproduct of strong people governance rather than the goal itself.
