From GIGO to Accountability: Data Quality, IRM, and AI

At Castlebridge’s Data Leaders’ Summit this year, I had a few minutes to speak as the new IRMS Ireland Group Chair. In a room full of data professionals, I wondered what I could meaningfully contribute.

I asked myself what I remembered from the first time I studied the technical aspects of computer science in the mid-to-late 1990s, when the internet was still in its infancy.

I realised that one principle from that time has stayed with me: GIGO – Garbage In, Garbage Out.

No explanations were necessary. Everyone in the room understood it immediately.

What was obvious then remains true now. Poor-quality data in leads to poor-quality results out. The difference today is not the principle – it is the scale and impact. In an AI-driven environment, GIGO is no longer just a technical truth. It is a business risk.

The Problem with Using GIGO as an Excuse

That point was sharpened later in the day when a panelist said he hated the phrase. My initial reaction was a quiet “ouch,” but his reasoning was worth paying attention to.

His issue was not the principle itself; it was how it is used. Too often, GIGO becomes a way to explain failure without taking responsibility for it. The data was bad, so the outcome is bad – and the conversation stops there.

That tension between recognising a problem and taking responsibility for fixing it, sits at the heart of what we see across organisations.

Why Data Quality Is a Critical AI Risk

AI systems do not just process data; they learn from it and scale it.

Yet many organisations are investing in AI while underestimating the condition of the data behind it. In our experience, most organisations do not believe they have significant data quality issues. Data appears to work well enough for day-to-day operations.

Problems tend to surface only under pressure such as during audits, transformation initiatives, or AI deployments.

Even when issues do appear, they are often absorbed into the “cost of doing business,” regardless of how high that cost becomes. Over time, this acceptance becomes embedded. GIGO stops being a warning and becomes an explanation.

The Real Challenge: Accountability for Poor Data

The real issue is not just poor data.

It is that organisations are not structured to fix it.

Without clear ownership, accountability and responsibility, data quality problems persist. Teams continue to work around issues rather than addressing their root causes, creating long-term operational and strategic risks.

Information and Records Management: The Foundation of Trusted AI

This is where Information and Records Management becomes critical.

It is often seen as a compliance exercise or a legacy function, but in practice it underpins everything AI depends on. It ensures information is reliable, complete, and usable over time.

Without that, data lacks context, outputs cannot be trusted, and decisions cannot be explained.

In an AI context, records are not passive. They are active inputs, providing the context and lineage that make outputs usable. Without well-managed records, AI operates without grounding, and trust erodes quickly.

How Data Governance Improves Data Quality

But understanding the problem is not enough. This is where data governance comes in.

Data governance is what turns awareness into action. It provides the structure that most organisations lack: clear ownership, shared standards, and defined accountability.

In practical terms, it is the glue that holds data quality improvement together.

Without governance, organisations remain reactive. Issues are identified and fixed in isolation, only to reappear elsewhere. With governance, data quality becomes continuous and managed. Problems are not just identified, they are addressed and they are prevented.

Data Quality Is a Human Issue, Not Just a Technical One

What becomes clear is that data quality is not a technical issue.

It is a human one.

It reflects how organisations define responsibility, design processes, and what they are willing to accept. If ownership is unclear, quality degrades. If issues are tolerated, they persist.

Technology can support improvement, but lasting change depends on people, behaviours and accountability.

Building Data and AI Capability Across the Organisation

This is why capability matters.

Organisations are investing heavily in AI, but often without a shared understanding of data or how to manage it. Training is treated as secondary, when in reality it is what enables governance to work.

Without it, accountability does not take hold and frameworks remain theoretical.

Moving from GIGO to Continuous Data Improvement

The shift required is straightforward, but not easy.

GIGO cannot be the end of the conversation. It has to be the starting point.

The challenge is not that organisations cannot recognise poor data. It is that they often choose to live with it.

In an AI-driven world, that is no longer sustainable.

The organisations that will succeed will not be those that adopt AI fastest, but those that take ownership of their data, put governance in place, and build the capability to improve it continuously.

Conclusion: GIGO Is No Longer an Explanation

GIGO still applies.

But it is no longer an explanation.

It is a signal.

Because once you recognise the garbage, the real question is whether you are prepared to fix it.