Looking Forward: From AI Hype to Human Reality

On December 16th, the Oireachtas Joint Committee on Artificial Intelligence published its First Interim Report, containing 85 recommendations for how Ireland should approach AI governance. In July 2026, Ireland will assume the EU Presidency and plans to host a summit that will shape Europe’s AI future. Between those two dates lies a narrow window that will determine whether Irish organisations lead or are led.

The question isn’t whether Ireland can lead on AI governance. We have the policy blueprint. There’s an implementation framework defined. We have world-class research. The question is whether Irish organisations have the governance maturity to leverage this advantage—or whether, when the tide goes out, we’ll find ourselves exposed on the beach.

The Convergence

Three things have come together at precisely the right moment.

First, the Oireachtas Committee has done something genuinely valuable: it has taken a first step to translate the abstract requirements of the EU AI Act into 85 actionable recommendations tailored to Irish circumstances. A National AI Office by August 2026. Mandatory algorithmic impact assessments for high-risk public sector AI. Annual evidence-based reporting. A Citizens’ Assembly on AI and digitalisation. These aren’t aspirations—they’re commitments against which Irish governance can be measured.

Second, Ireland’s Data Governance Roadmap, developed through the EMPOWER programme at the Innovation Value Institute in Maynooth University, provides the implementation framework that most countries lack. The Roadmap takes a three-tier approach—national, organisational, and individual—with four macro-level themes spanning leadership, policy, skills, and technology. It offers a maturity model that moves from “absent” through to “advancing,” giving organisations a clear path from wherever they are to where they need to be. (disclosure: I was one of the team who developed this framework)

Third, we have timing. Ireland’s EU Presidency in July 2026 creates a platform that few other member states will have. We get to showcase our approach while the world is watching. That’s an extraordinary opportunity—if we’re ready to seize it.

What Lies Beneath

Here’s the uncomfortable reality that the AI hype cycle has been obscuring: you cannot build transparent AI systems on top of opaque data practices.

The Oireachtas Committee requires that “accountability, traceability, transparency and explainability must be embedded in AI.” That’s Recommendation 45, and it’s entirely sensible. But here’s the paradox: many Irish organisations demanding AI transparency cannot answer basic questions about their existing data.

Where did this data come from? What does it actually mean? Can we trust it? Who is accountable for it?

If you can’t explain your data, you cannot explain your AI. The transparency requirements proposed by the Committee will expose what lies beneath—and for too many organisations, what lies beneath is a legacy of “data debt” accumulated over years of treating data management as someone else’s problem.

The Data Governance Roadmap is blunt about this: “Absent or ineffective governance prevents value extraction and can turn a valuable asset into an expensive liability.”

This isn’t theoretical. When algorithmic impact assessments become mandatory, organisations without data lineage won’t be able to assess bias in their training data. When transparency mechanisms are demanded, organisations without metadata standards won’t be able to explain what their AI systems are actually doing.

The hype told us AI would transform everything. The reality is that AI reveals everything—including the governance gaps we’ve been pretending don’t exist.

The Leadership Gap

The Roadmap research found something that will be familiar to anyone who has worked in data governance: “Leadership is often dependent on individuals’ best efforts rather than systematic structures.”

We’ve all seen this. Organisations where data governance lives or dies based on whether one passionate advocate remains employed. Boards that can discuss financial risk in granular detail but go glassy-eyed when data risk is mentioned. Leadership teams who approved AI pilots without ever asking how the data was sourced, governed, or quality-assured.

The National AI Office highlighted by the Committee addresses the national leadership vacuum with a structural governance element. But it doesn’t, and it can’t, address the organisational leadership gap. That’s on us.

At the Data Leaders Summit over the past two years, we’ve been tracking this theme. In 2024, we discussed how board-level data acumen would become as essential as financial literacy. In 2025, we examined how AI governance must be a subset of data governance, not a separate discipline. The pattern is clear: organisations that treat AI as a technology project rather than a governance challenge are setting themselves up to fail.

Delegates at the 2026 Data Leaders’ Summit next March will tackle (amongst other things) a pointed question: how many Irish organisations have genuine board-level AI and data governance capability? And how many are still depending on individual champions who might move on next month?

The Skills Transformation

The Oireachtas Committee’s ambition is laudable: “a coordinated national effort across education, from primary school to workplace” (Recommendation 28) to support AI literacy. But we need to be careful about what “AI literacy” actually means. Recommendation 28 of the Oireachtas Committee echoes one of the key recommendations in the Data Governance Roadmap, which recommends the development of foundational skills for data governance and data acumen from primary education at the “Enhancing Level”

At the 2024 Summit, we introduced the distinction between Data Nous, Data Literacy, and Data Acumen. Nous is about intelligence of purpose, having ‘common sense’ about data. Literacy is about technical skills and knowing how to work with data tools. Acumen is about contextual understanding, such as knowing when and why to apply different data management capabilities.

Too often, “digital literacy” programmes have focused exclusively on tool operation while ignoring critical thinking. We’ve taught people how to use spreadsheets without teaching them how to question the data in those spreadsheets. We’ve trained people on AI prompting without helping them understand what the AI is actually doing with their data.

The risk is that we repeat this mistake with AI literacy. Training people to use AI tools is the easy part. Developing the critical thinking to understand AI limitations, to question AI outputs, to recognise when AI is inappropriate—that’s the hard part. And it’s the part that actually matters.

The National Data Governance Roadmap is clear that skills development has been ad hoc and conflated with technical skills, particularly at the “Initiating” level of maturity. The Oireachtas Committee wants AI literacy from primary school to workplace. AI Literacy needs to be built on data acumen.

The gap between aspiration and reality is where the actual work needs to happen.

The Democratic Question

There’s a recommendation in the Oireachtas Committee report that deserves more attention than it’s likely to get: the Citizens’ Assembly on AI, Digitalisation and Technology.

This is genuinely significant. It acknowledges that AI governance isn’t just a technical or commercial question—it’s a democratic and societal one. The decisions we make about AI will affect everyone, including those who have no voice in technology procurement or policy development. (This theme is discussed by my co-author Katherine O’Keefe and I in the 2nd Edition of our book “Data Ethics“).

But here’s the challenge: meaningful democratic participation requires understanding. If citizens can’t engage substantively with AI governance questions, then “participation” becomes theatre, descending into a box-ticking exercise that legitimises decisions already made by technical elites.

The Committee also recommends ensuring that the voice of underprivileged communities is not filtered out by AI. Recommendation 40 requires a key focus on avoiding “deepening existing inequalities or worsening the digital divide”. It raises a profound question: how do we build AI governance that genuinely includes everyone, not just data professionals and technology leaders?

At the Data Leaders’ Summit, we’ve always taken the view, as articulated by John Ladley, that data is anthropological—it’s the “24/7 digital record of human activity”. It affects people directly and indirectly, whether or not they’re aware of it. That perspective becomes even more critical as AI amplifies data’s reach and impact.

The Business Case

For those who need commercial justification (and most boards do) here’s the straightforward case.

Organisations that invest in data governance fundamentals now will be positioned to deploy AI with confidence, meet transparency requirements without panic, and demonstrate the accountability that regulators and customers increasingly demand. They’ll be able to engage constructively with the National AI Office when it launches. They will be able to hold their heads up high during Ireland’s EU Presidency as paragons of the possible.

Organisations that don’t make these investments will face a different trajectory. Regulatory scrutiny. Reputational damage. High-risk AI systems that can’t be deployed because impact assessments reveal ungovernable data. Public accountability failures when annual reporting exposes the gap between AI ambitions and data reality.

The 2024 Summit concluded that “the challenges of today are often the result of compromises of the past.” The corollary is that the advantages of 2027 and beyond will be built on the investments of 2026. Organisations that delay will find themselves trying to build governance capability while their competitors are already extracting value.

The Path Forward

What does this mean in practice?

Organisations should be assessing their data governance maturity maturity against some form of objective criteria. Not as a theoretical exercise, but as an honest reckoning with what lies beneath. Where are the governance gaps? Where is the data debt? What would an algorithmic impact assessment actually reveal?

The March 2026 Data Leaders Summit at the Ferrycarrig Hotel in Wexford will be a useful checkpoint in this critical period. It’s where Irish and international data leaders get together in a ‘demilitarised zone’ to share what’s working and what isn’t, identify common barriers, and align organisational readiness with national ambitions, and network. 2024 and 2025 discussed the importance of foundational issues in data management as enablers for compliance, environmental responsibility, ethics, and AI implementation. In 2026 we carry that theme further.

The Changing Tide

The theme for the 2026 Data Leaders’ Summit is “The Changing Tide.” It captures Ireland’s moment: the surging tide of AI opportunity, but also the risk that when the tide goes out, it will expose those who haven’t built on solid foundations.

Most countries have policy or framework. Ireland has both. Most organisations have AI ambitions. The question is whether they have the governance maturity to achieve them.

The Oireachtas Committee has charted the course. The Data Governance Roadmap provides the navigation tools. What happens next depends on whether Irish data leaders are prepared to do the work needed to address what lies beneath, to close the leadership gap, to build genuine capability rather than governance theatre.

The tide is changing. The question is whether Irish organisations will ride it or be swept away by it.

---

The Data Leaders Summit 2026 takes place on March 18th at the Ferrycarrig Hotel, Wexford. Early bird registration is now open at dataleadersummit.ie. For organisations seeking support with data governance, AI readiness assessment, or leadership development, contact Castlebridge to discuss our advisory and training services for 2026.