Shaping AI: What should it do and for whom?
28 April 2026, Dublin, Ireland: A new NESC report Artificial Intelligence in Service of Society: Navigating Our Way Forward emphasises that Ireland is currently in a critical window of opportunity. As AI becomes increasingly embedded across public services, workplaces and everyday life, now is the time to put the right foundations in place: strengthening skills, governance, infrastructure and public trust so that Ireland can realise the benefits of AI while minimising foreseeable risks and unintended consequences.
The report outlines that the goal should be to proactively shape AI, so that its benefits can be realised responsibly, equitably, and sustainably. Taking a broad socio-technical perspective, the report argues that AI is not merely a technological tool, but a transformation shaped by governance, institutional capacity and societal choices. According to Dr Siobhán O’Sullivan, Senior Policy Analyst at NESC:
“There is a tendency to treat AI as a purely technical phenomenon—something to be evaluated on the basis of whether it works as designed. But that framing misses the most important questions. AI systems do not operate in a vacuum; they are embedded in organisations, workplaces, public services and communities, and their impacts emerge from that interaction. A socio-technical lens asks not only whether a system functions as intended, but who benefits, under what conditions, and at what cost. That shift in perspective is what moves us from asking what AI can do, to asking what it should do—and for whom.”
The report sets out five interconnected priorities:
- Responsible and Strategic Adoption: AI should address clearly defined public and organisational needs and align with workforce skills, data quality and institutional capacity.
- Trustworthy and Ethical Practice: Systems must be transparent, accountable and subject to meaningful human oversight, translating ethical principles into real-world actions.
- Anticipatory Governance: With AI evolving rapidly, regulation must be forward-looking, adaptive and based on continuous monitoring rather than reactive fixes.
- AI Literacy as National Infrastructure: Building widespread understanding of AI is essential for workforce adaptation, democratic oversight and responsible use.
- Public Legitimacy: Long-term success depends on securing public trust through inclusive engagement and sustained societal dialogue.
A central theme of the report is the requirement for AI systems to be safe, ethical and trustworthy in practice, not only in principle. AI systems are probabilistic and imperfect. Meaningful human control is essential to prevent over-reliance, loss of judgement and accountability gaps. High-level ethical principles must be translated into concrete practices with individuals and institutions building genuine ethical capability to ensure AI operates safely, fairly and effectively.
Among the report’s most significant findings is its designation of AI literacy as essential national infrastructure. Ireland has a growing ecosystem of AI literacy initiatives, but these remain fragmented, and significant gaps persist in public understanding. NESC calls for development of AI-literate citizens capable of questioning and scrutinising AI systems, and AI-literate senior leaders capable of providing effective organisational oversight—both of which are identified as preconditions for democratic accountability.
On governance, the report notes that the trajectory of AI capability remains genuinely uncertain, and that regulatory approaches must be agile, anticipatory and continuously updated. In line with the National Digital and AI Strategy, NESC argues that continuous monitoring is central—particularly given evidence gaps and the tendency for AI systems to behave differently in complex real-world environments than in controlled settings. Anticipatory governance enables policymakers to detect emerging risks early and respond proactively rather than reactively.
Dr Larry O’Connell, Director of NESC, noted that:
“The trajectory of AI capability is genuinely uncertain. We cannot predict with confidence what the technology will look like in five or ten years, which means governance designed only for today’s systems may be inadequate for tomorrows. Anticipatory governance gives us the tools to prepare for multiple possible futures — through strategic foresight, continuous monitoring and flexible regulatory approaches that can detect emerging risks early and respond proportionately. The goal is not to predict the future but to build the institutional resilience to navigate it, whatever form it takes.”
To read the report in full please click here.
ENDS
For further information, contact:
Marie Lynch, Carr Communications,
e: marie@carrcommunications.ie | m: 087 973 0522
