Regulation of Artificial Intelligence Bill 2026: Second Stage Seanad Éireann — 2026-07-01 ============================================================ Patricia Stephenson (SD), Labour Panel Al regulation has been a long time coming. It is one of the most pressing challenges we face. It has changed the way we work, our entire understanding of the world and our ability to discern truth from fiction when we see what AI can produce. It is good we are discussing regulation and this directive is finally being transposed but I want to express serious concern about the manner in which this is being done. It is being rushed through with very little parliamentary scrutiny, despite our having years of lead time to prepare for this. While preparing my notes for this evening, I could not even read back the transcript for last night, such is the speed with which we are rushing this through at the end of term and not allowing meaningful time for scrutiny of such a large, complex piece of legislation. The establishment of a statutory body and the regulatory framework for the supervision and enforcement of the Al Act in the State is inherently complex. It is critical we get it right and we have to have sufficient time to do that, which was not afforded to us. I echo the comments of my colleague, Deputy Gibney, in the Dáil last night that it is troubling that the Al committee did not have the opportunity to engage effectively with this Bill, despite the gravity of what we are dealing with. While I welcome the move to give full effect to the EU Al Act, as provided for in this Bill, we need to be treating this as the “minimum baseline for national Al regulation, not a maximum standard”, as was the view of the joint AI committee. Before addressing the content of the Bill, I want to speak to larger concerns I have with the Government's approach to AI and big tech regulation. Time and again we hear the industry narrative being peddled that we need to maintain the competitiveness and viability of Ireland's digital economy. The conversation is always framed in terms of innovation and enterprise, rather than the protection of fundamental rights and the litany of harms and threats posed by this tech, which is evolving faster than many governments in Europe, including this one, can keep up. The location of the new Al office within the Department of enterprise speaks to the centrality of this business-first approach to governance, rather than putting human rights first. It further speaks to the power of the big tech lobby at home and in Brussels, which continues to consolidate its power and influence over politics to a frightening degree. We see this with passage of the EU’s digital omnibus on Al - a worrying deregulation agenda. I know the Minister of State, Deputy Smyth, does not agree with me on that but it is serving to critically weaken the Al Act and playing right into the hands of big tech. We had some wins in provisions around data participation in the digital omnibus but simplification is weakening rights, rather than strengthening them. There will still be use of incredibly sensitive data in AI systems and the high-risk implementation period has been extended. High-risk systems do not need to be compliant until next year. AI companies will rush out systems that can have incredibly dangerous implications. It is ridiculous that they will have this window in which they can still produce this because the high-risk system regulation will not be retroactive for new pieces of tech. That is stuff that might be dealing with complex and sensitive child welfare cases and medical AI cases. We know, in the case of social media and entertainment, that AI is algorithmically driven to addict the human brain. All of those high-risk systems have this extra time in which they can be developed and not come under the regulation. That is really worrying. We can talk about the enormous opportunity and potential for good, regulated AI. I see it particularly when we talk about medical science, but we also need to get serious about the scale of harm and potential for harm coming down the road to us if we do not regulate these companies in any meaningful way. The Grok AI nudification scandal showcased how completely unprepared we are at home and abroad, at EU level, to counter these threats and X effectively got away with a slap on the wrist. The company was asked to please not do it again and allow sexual abuse imagery material of children and adults to be developed on its platform because we would really like it not to. It shows the complete contempt with which these companies view us and governments. They did not even attend the Oireachtas media committee, such is the contempt with which they view our parliamentary process. We have seen no prosecutions and no sanctions for this, and the stark reality is that this feature still exists and can be accessed in this country with the right technology. We have not effectively safeguarded the public, including women, children and minorities, from this harmful technology. It is still accessible. People can still promote child sexual abuse material and image-based sexual material using the technology made available to them by a private company that operates in Ireland. Another threat we cannot overlook is that of labour displacement. Recently, 720 workers from the Dublin-based Covalen company heard they were going to be made redundant thanks to AI. This is just coming down the line to us. My Dáil colleague Deputy Gibney highlighted that the Government's stock response to this threat from AI is to hail the potential for AI to augment the quality of jobs by performing more menial tasks. That is not the reality of what is happening. Automation has not served to improve the quality of jobs people are doing, job security, job stability or the pay and conditions of workers. People are losing jobs from AI and automation. Customer support is all bots now. There used to be humans doing these jobs. They are now done by bots. We are going to see a fundamental shift in society because of AI labour displacement. Gen Z is struggling to find jobs as companies pivot to AI, using it for entry-level jobs. Youth unemployment is worrying in Ireland, but across Europe and the world youth unemployment is skyrocketing and on the rise. That is because companies see it as financially more beneficial to use an AI system they do not have to pay than upskilling a young worker. We are not facing up to the reality of this by having any meaningful strategic planning in place. I turn to the Bill itself. Serious concerns have been raised in the Dáil regarding the governance and potential independence of the new AI office. It is vital that the AI office as a regulator has the appropriate independence and resources to operate to ensure it can carry out its functions protected from potential corporate influence. The AI office is a vital piece of our new regulatory framework and must have sufficient independence from Government if it is going to be effective. Yet, there currently real deficiencies in how the Bill provides for this crucial independence. There are concerns regarding transparency of appointments to the board and of its leadership. The Bill currently suggests that the first board and CEO will be directly appointed by the Minister for enterprise with little reference to any binding recommendations from the Public Appointments Service or procedures in the appointment to safeguard independence. Without independence it is actually kind of meaningless. The practice we see in the legislation does not tally with Government's assurances as to its independence, nor does the provision where the Minister can instruct the office to act on his or her direction, even with the exceptions for specific enforcement matters. This is not any particular Minister. I am talking about a Minister or a Department. Undue influence from the Minister will undermine the effectiveness of the regulatory system. How can we have an effective regulation of AI in this country when the supposedly independent State body tasked with that regulation is not permitted to comment on Government policy when asked by Members of these Houses at Oireachtas committees? I am also concerned that a key aspect of the EU directive is missing from this transposition. It is the part about vulnerable groups seeking an explanation as to why AI tools have made a specific decision. I might be wrong and perhaps it is in the legislation, but I do not think it is. There have been multiple examples of algorithmic AI decisions being made in EU countries where they are having a biased reading of the applicants, particularly in the Netherlands with child welfare cases. If we do not have a mechanism by which vulnerable groups can seek to understand an explanation for an AI decision in an automated process at government level or in the private sector, we are serving to emphasise the further marginalisation of those communities. That is really crucial. It may be in the Bill, and I did not see it, but that just points to the fact that we have not had proper time to meaningfully scrutinise this legislation. If we want to talk about being at the centre of EU excellence and digital regulation, we have to take seriously meeting our obligations under the AI Act by establishing a truly independent and robust regulatory and enforcement regime. --- Source: Houses of the Oireachtas. Licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). The Official Report is revised after first publication; the fetch timestamp below identifies the version quoted. Record URI: https://data.oireachtas.ie/akn/ie/debateRecord/seanad/2026-07-01/debate/main Retrieved: 2026-08-14T04:55:08+00:00 Sitting date: 2026-07-01