Albert Dolan questions the clarity of the AI Bill's regulatory model and urges stronger cooperation between existing regulators, arguing businesses need a single, understandable point of engagement. He also stresses the importance of regulatory sandboxes for SMEs and wider investment in AI literacy so Ireland can protect citizens without stifling innovation.
Regulatory clarity needed
Albert Dolan highlights concerns that multiple bodies-from Commission na mBan and the Data Protection Commission to the Workplace Relations Commission and Revenue-could leave firms facing overlapping or unclear oversight. He argues the proposed distributed model must be refined so regulated entities know which regulator to approach and how cooperation between regulators will be formalised.
One-door cooperation and harmonised powers
Dolan says cooperation mechanisms in the AI Bill and related data legislation should operate like a one-door policy, giving businesses a clear route to engage. He recommends harmonising supervisory powers and making the regulators' cooperation more formal to avoid gaps where no regulator takes responsibility.
Support for SMEs and sandboxes
Dolan stresses that the regulatory sandbox will be essential for SMEs and larger enterprises to field-test AI innovations in a safe environment. He argues sandboxes, paired with active policy measures, help firms adapt while keeping citizens protected.
AI literacy and skills investment
Beyond legislation, Dolan calls for investment in AI literacy across the economy, echoing the idea of AI literacy as infrastructure. Building skills and resilience, he says, is crucial to safeguarding people while accelerating responsible AI adoption in Ireland.
We publish thousands of recordings to make Irish politics transparent and resistant to manipulation. Spotted an error? Report it — together we are building a reliable archive of Irish politics.
Thank you both Eric and Áine for being with us today and I guess just really to discuss the merits of this AI Bill and maybe where you see things that could change or could be improved and what the general feedback is on it to this point and the first question really is around this regulatory model is it clear enough for businesses or is there a fear that firms will be left dealing with overlapping regulators and I guess by that I mean like you know you have your Commission na mBan, you have your Data Protection Commission, you have the Workplace Relations Commission, you have revenue, you have so many different bodies already in the state to deal with. I guess what we would hate to see is the AI Bill essentially creating another regulator, albeit it's vital that we have the expertise to handle AI related issues that will arise no doubt but I guess the question is, is this model clear enough for your members? Should we take a few questions or just... No, fire it with that one first. So I suppose first of all one of the areas where we feel that refinements could be made is it is in this area of coherence so like we fully understand the idea of a distributed model okay so the logic behind the distributed model is that there's domain knowledge that already exists because if you think about the AI Act it involves its product safety as well as protecting fundamental rights so you have existing regulators in market surveillance so the logic is to I suppose leverage that domain that regulatory domain knowledge so we acknowledge that we accept it but I think what we point to in our you know our written submission that we've given and also in our opening statement is that there's areas where that could be actually refined so just like if we go through maybe a few of the areas where we feel it could be refined. I suppose first thing is there is this there is a mechanism in the data bill for instance about cooperation forum there is also a mechanism in the AI Act for a cooperation mechanism as well but that's for the regulators to talk to each other what we're recommending is that should be a one-door policy like there's doesn't seem very clear to us that we could actually engage with that that cooperation forum so therefore to answer your question it's not clear to us okay and we do feel that a mechanism by which it's a it's a kind of one door that we we go to to get the answers is it another point just just on that you feel that you know the likes of IBEC or other entities should have a seat at the table on that forum with the regulators or like you know could say for example like leaf you know you have a seat at leaf and that gives you a position to engage from yeah are you saying that there's a fear that you don't have a mechanism to engage the fear is for the people who are actually obligated or regulated who do they who do they talk to it's not clear for from from reading you know the proposal saying in the case of AI which regulator takes a case you know and and then what happens if two regulators don't decide to take a case and so on right so there's a few areas that we point to in our submission where there's a bit more certainty needs to be brought in what we're really saying here is these cooperation mechanisms so on need to be more formalized I think is what we're trying to say and also we point to the basis of the powers right in the law so you have existing regulators and they're going to actually depend on existing powers we're not sure whether those existing powers kind of exceed what's in the AI Act and market surveillance okay so and that's why we're also saying that supervise supervisory power should be harmonized I see the key takeaway that I suppose we wouldn't leave with the committee is is this harmonization idea we're not against the idea that there could be 13 regulators involved we just want to make sure that the mechanisms in this bill there's mechanisms in this bill that it's very clear they're talking to each other and they're talking to those who are actually obligated it's less of a it's a lot less of a concern that IBEX at the table it's more concern for those who are regulated because you could have several different regulators which regulator do you go to is the question and so I guess then on the other side of it is say for companies that you represent and particularly SMEs yeah what compliance risks I guess and what practical supports should government put in place to help firms adapt yeah and ultimately mitigate those risks okay well in the actual bill and what we're talking about here is more to do with the actual regulatory sandbox okay so the sandbox is going to be important I think from the point of view it allows an SME but we also think it should be broader yeah right to I suppose field test an AI idea or innovation in a safe environment because the idea of regulatory sandboxes and I think they these exist already in certain sectors is they're kind of working with the regulator in field testing that basically that it's AI act ready the innovation that they come out with so so for us I think the regulatory sandbox is going to be really really important for SMEs but also enterprise in general you have to think of AI as not just kind of a tech issue it's going to be a horizontal business issue because AI is a horizontal technology and it's going to be used by a lot of different sectors and just my understanding would be that the I think the central bank has an active sandbox already and so that's I guess an example where you know fintech can flourish without the fear of being reprimanded and so I guess my last question because I'm conscious of the time sure would be you know there's always this opposing narrative of you know we need to be very careful with AI and that we you know we need to protect our citizens and that is always at the forefront of our mind as legislators is to make sure that we protect our citizens we put our citizens first how can we make sure that we do that without slowing down AI because I guess Ireland is a place that has benefited greatly from innovation and technology over the last many years how do we make sure that we can have a you know a collaboration between the two that we can protect people but we can still technology over the last many years how do we make sure that we can have a you know a collaboration between the two that we can protect people but we can still accelerate the rate at which we adopt AI and utilize it to its full potential. Well that's the second part actually because it's not just going to be sandboxes that are necessary it's actually a policy action as well so it's not just legislative and I think part of this is really it's to do with literacy AI literacy across the population across the economy I think Nesk have called it infrastructure AI literacy as as infrastructure and we're a modern economy so I think investment in the right skills is going to be really key to making everybody resilient so that you safeguard people but also encourage the innovation.
Thank you for downloading 🙏
If you publish this material on social media, we would be very grateful if you tagged VideoParliament. It helps us reach more people and keep building a transparent archive of Irish politics.
