Ireland Publishes AI Bill 2026 to Establish National AI Office

Dublin Unveils Landmark Bill for National AI Watchdog

The Irish government has unveiled the blueprint for its domestic regulation of artificial intelligence, publishing the General Scheme of the Regulation of Artificial Intelligence Bill 2026. This landmark legislation provides the first detailed look at how Ireland intends to implement the European Union’s sweeping AI Act, establishing a new national supervisory body and a robust enforcement framework designed to balance innovation with public safety.

The Bill solidifies Ireland’s strategy to operate a ‘federated’ model of regulation, leveraging the expertise of existing sectoral regulators under the stewardship of a newly created central authority. This move positions Ireland, a major European hub for global technology firms, at the forefront of AI governance, aiming to create a trusted environment for the development and deployment of AI systems while navigating the complex challenges posed by this transformative technology.

Setting the ‘Golden Standard’ in a New Regulatory Era

The Irish legislation is a direct consequence of the EU AI Act, a pioneering piece of regulation that sets a global benchmark for AI safety. The EU’s approach is fundamentally risk-based, treating AI not as a monolithic concept but as a spectrum of technologies with varying potential for harm. It regulates AI systems much like other products, imposing stringent pre-market obligations on those deemed ‘high-risk’, such as systems used in critical infrastructure, medical devices, or law enforcement. This framework is designed to build public trust and encourage capital investment into a safe and predictable European AI market, attracting capital from insurers, pension funds, and public buyers who prioritise lower-risk deployments.

For technology companies, particularly those from the United States with significant operations in Ireland, this represents a pivotal shift. The compliance conversation is moving beyond the data privacy and cybersecurity remits of GDPR and NIS2 towards a more demanding focus on product safety and compliance engineering. While some have viewed the EU’s assertive regulatory stance as a form of digital protectionism, proponents argue it is a jurisdictionally neutral market hurdle designed to elevate standards globally. The goal is to create a ‘golden standard’ for safe AI, ensuring that systems placed on the EU market are secure, transparent, and respect fundamental rights, regardless of where they were developed.

Introducing the AI Office of Ireland

At the heart of Ireland’s new legislative scheme is the establishment of the Oifig IS na hÉireann, or the AI Office of Ireland. This new statutory body is tasked with the monumental responsibility of overseeing the AI landscape. The Bill designates the Office as the lead Market Surveillance Authority (MSA) and the national Single Point of Contact, making it the central coordinator for all AI-related regulatory matters in the state.

The Office’s structure is designed to ensure independence and robust governance. It will be led by a Chief Executive Officer who reports to an independent, seven-member board. Its core functions are multifaceted: it will facilitate the consistent enforcement of the AI Act across all sectors, foster seamless cooperation between the various competent authorities, and act as a hub of technical expertise. Crucially, it is also charged with a proactive mission to promote AI innovation and adoption, alongside a vital public-facing role in fostering greater AI literacy among citizens and businesses alike.

Balancing Innovation and Oversight

A key feature of the proposed legislation is its dual focus on both regulation and enablement. The AI Office will be responsible for establishing and maintaining a National Register, a public-facing database that will list all instances of prohibited AI practices as defined under Article 5 of the EU Act. It will also catalogue all high-risk AI systems in use, providing an unprecedented level of transparency and public scrutiny.

Simultaneously, the Bill champions innovation through the creation of national regulatory sandboxes. These controlled testing environments are seen as crucial for the safe development of cutting-edge AI. They will allow developers, from startups to established enterprises, to experiment with new systems under the guidance of regulators, identifying and mitigating risks before a product is brought to market. In a significant boost for the domestic tech ecosystem, the legislation mandates that small and medium-sized enterprises (SMEs) be given priority access to these sandboxes, lowering the barrier to entry for smaller players and fostering a more competitive market.

A Measured Approach to Enforcement

The Bill provides much-needed clarity on the enforcement powers of the various authorities. It formally designates several existing bodies as Market Surveillance Authorities within their respective domains, clarifying their operational relationship with each other and the central AI Office. This distributed model ensures that deep sectoral knowledge is applied to AI oversight, whether in finance, healthcare, or transport.

Significantly, the legislation builds in a powerful ‘checks and balances’ system by requiring High Court approval for the most stringent enforcement actions. According to the scheme, an MSA seeking to compel an entity to comply with the rules must first apply to the High Court for an order. Similarly, any administrative sanctions or fines imposed by an adjudicator will only take effect once confirmed by the High Court. This two-step process is designed to provide legal certainty, protect against overreach, and ensure that penalties are applied fairly and proportionately, a crucial consideration for businesses operating under the new regime.

An Evolving Digital Landscape

One area of slight uncertainty remains the potential impact of the EU’s proposed ‘Digital Omnibus’ regulation. This separate legislative initiative aims to streamline enforcement procedures across several key digital laws, including GDPR, the Digital Services Act, and the Digital Markets Act. The annex to the Irish Bill acknowledges that these proposals will continue to inform the national implementation of the AI Act. However, the government has stressed that the core deadline for establishing the AI Office of Ireland by 2 August 2026 remains firm. Should the Digital Omnibus, once finalised at the EU level, require amendments to the Irish law, the government has committed to considering them, signalling a flexible approach to an ever-evolving digital regulatory environment.