ICBE Webinar: Building AI Literacy – Practical Skills and Governance / From Awareness to Action

EU AI Act Deadlines Loom

The European Union's Artificial Intelligence Act, which entered into force in August 2024, imposes a risk-based framework on AI deployment. Prohibitions on unacceptable-risk practices and basic AI literacy obligations for staff took effect in February 2025. The most substantial requirements—for high-risk AI systems in areas such as hiring tools, credit scoring, biometric identification, and critical infrastructure—apply from August 2026, with some extensions to 2027 for legacy systems.

Companies operating in or selling into the EU must now prepare for conformity assessments, transparency rules, and ongoing monitoring. The stakes are concrete: fines for serious breaches reach €35 million or 7% of worldwide annual turnover, whichever is higher. Even partial non-compliance in high-risk deployments can halt product launches or trigger market withdrawals, especially for multinationals with European exposure.

Ireland, home to many global tech firms, is aligning through the Regulation of Artificial Intelligence Bill 2026 and a new AI Office of Ireland, set to coordinate enforcement by August 2026. This office will also promote adoption and literacy, reflecting a dual aim to foster innovation while enforcing rules. Recent national strategies emphasize skilling and sandboxes for testing, yet the distributed regulatory model—relying on sectoral authorities—risks inconsistencies or delays.

A key tension lies in balancing compliance costs against competitive pressures. Smaller firms and startups face disproportionate burdens in upskilling workforces and documenting literacy efforts, while larger players may leverage resources but still grapple with interpreting vague literacy standards. Delays in EU guidance, such as missed deadlines for high-risk clarifications, compound uncertainty, pushing businesses toward precautionary over-investment in training and governance.