A draft strategy published by the Executive Office proposes using Artificial Intelligence across Northern Ireland's public sector to handle routine administrative work such as document processing, data entry, minute-taking and basic query handling. The document does not propose job losses but signals a shift in how civil servants carry out their duties, with officials expected to spend less time on repetitive tasks.
According to the draft, staff would be:
"freed up for human interactions by automating routine public-facing tasks"
The paper, produced jointly by the Executive Office and the Department for the Economy, also states that:
"basic technologies should be rolled out as a priority to increase efficiency in routine office tasks"
What other services could be affected?
Beyond office administration, the draft flags several other functions as candidates for AI-driven change. It suggests AI could accelerate public consultations to allow, in the document's words, a "swifter passage to policy", and could also speed up the processing of planning applications and the analysis of large public datasets used to inform policy decisions.
How would the transformation be managed?
The draft acknowledges that attempting sweeping change all at once would not work, stating that:
"large-scale transformation in one go is virtually impossible"
Instead, it recommends breaking the process into smaller initiatives described as "readily comprehensible projects with a clear emphasis on value and impact", supported by staff training and dedicated space for testing new approaches. The document also proposes creating "human oversight teams" within each department, made up of staff with technical know-how who also act as "critical thinkers and challengers" of AI-driven decisions. Human oversight is listed among eight principles in the consultation, which the document says "sets an operational expectation for how AI should be developed, procured, and managed".
Will Northern Ireland build its own AI systems?
Rather than developing bespoke AI tools from the ground up, the strategy favours what it calls a "smart second-mover" approach — adapting off-the-shelf technologies that have already proven successful elsewhere. This mirrors recommendations in a separate public-sector AI report, which argues for shared, reusable AI components for common service patterns that could be applied consistently across departments rather than being built individually each time.
The same report also recommends a mixed infrastructure strategy combining public cloud services with private AI infrastructure for sensitive services, an approach that could inform how Stormont balances cost, security and computing capacity as it moves forward.
Why does the strategy call for more data centres?
The document argues that Northern Ireland needs to expand its computing infrastructure — the data centres that underpin AI systems — to reduce dependence on external cloud providers. It states that the region should develop:
"the necessary compute infrastructure to avoid vendor lock-in with cloud services"
In practice, this means building local capacity rather than relying solely on computing power purchased from firms such as Amazon Web Services. The draft adds:
"Cloud is too expensive for training new AI models, for allowing space for experimentation, or for deep R&D. Sovereign, green, high-performance computing is therefore essential"
Data centres consume large amounts of electricity and, depending on design, significant volumes of water for cooling. In the Republic of Ireland, such facilities now account for a quarter of all electricity consumption, a trend that has become politically contentious.
How does this fit with wider public-sector AI plans?
The proposals follow the creation of Northern Ireland's Office of AI and Digital, announced in the February Programme for Government, which was given responsibility for coordinating digital transformation with NISRA and other digital leaders across the civil service. According to PublicTechnology, the office will support a forthcoming AI strategy and action plan, with chief scientific and technology adviser Professor Helen McCarthy leading a roadmap for AI adoption across departments.
Other administrations are pursuing similar goals. Scotland's newly established AI agency has set a target of contributing £23bn to its economy by 2035 through AI-driven improvements to public services including healthcare, while also addressing ethical and environmental concerns. Analysis from McKinsey suggests public-sector AI strategies generally work best when they focus on resident-facing outcomes, rework entire workflows rather than isolated tasks, and keep humans involved in consequential decisions — a principle reflected in Stormont's proposed oversight teams.
What happens next?
Northern Ireland's full AI strategy and action plan is expected to be developed under the Office of AI and Digital, with Professor Helen McCarthy continuing to lead the roadmap for AI adoption across government departments, according to PublicTechnology. The current document remains a draft strategy open for consultation before any formal adoption.
Key Facts
- The draft strategy was published by the Executive Office alongside the Department for the Economy.
- Targeted areas for automation include document processing, data entry, minute-taking and query handling.
- The document proposes eight guiding principles, including mandatory human oversight of AI systems.
- It calls for expanded local data centre infrastructure to reduce reliance on external cloud providers.
- Professor Helen McCarthy is leading the wider AI adoption roadmap under the Office of AI and Digital.
