Dail Éireann · 2024-01-18
Part of Other Questions
Party shown is the one held on 2024-01-18, frozen at parse time — not the speaker's party today.
7. Deputy Aindrias Moynihan asked the Minister for Health in what areas in acute hospital settings AI technology is being considered; and if he will make a statement on the matter. [2032/24]
One of the big challenges is that security and confidentiality will be needed, and buy-in and consent from patients for their data to be used in that manner. It trialled the use of AI software and looked at 15,000 patient scans to assess the potential to speed up care in the emergency department.
If we are looking at having a huge store of people's data and running AI over it to predict where different populations might be impacted by various conditions, we need buy-in from people, trust in AI and consent for the use of their data. There are areas we are looking at very closely such as cybersecurity, as we discussed, ethical AI and the guidelines on the use of that, patient consent, and full information on how information may be used.
Built by scoring every sentence in the debate on how distinctive its language is, then quoting the highest-scoring ones back word for word and in the order spoken. Nothing is paraphrased or invented — if a sentence appears here, it was said in the chamber. Drawn from 927 words across 6 contributions by 2 speakers.
How often each word appears across the whole debate, ranked by how distinctive it is to this debate rather than common to all parliamentary language.
All 6 contributions, start to finish, in the order they were made. The one you came here for is highlighted.
Gabhaim buíochas leis an Leas-Cheann Comhairle. Athbhlian faoi mhaise di. The use of artificial intelligence in healthcare settings can offer very exciting possibilities, and also present challenges and various considerations. Will the Minister give an outline of the use of AI in medical settings throughout the country, as well as the various considerations he has had of its use?
I thank the Deputy for the question. I heard this morning that AI is front and centre at Davos as well. It is certainly coming to the fore very rapidly. It will have potentially radical benefits, be they in diagnostics, in some advanced care, or robotics for procedures, and as we move forward with genetics, genomics and genome sequencing. AI will move more and more to the fore. What is critical for Ireland is for us to make sure that we capture these benefits for patients. We must have state-of-the-art e-health and digital infrastructure in place. The reality is Ireland is not just a laggard but a significant laggard in the EU and the developed world in this regard.
To that end, we have been doing a lot of work on a new e-health and digital strategy. I will launch that strategy this year. It is being finalised at present. It will be a six- or seven-year strategy. We have had a lot of consultation with experts and potential providers to see what platforms can be used. We are doing it in parallel with the health information Bill, which I will bring to the Government shortly and will include the matter of unique patient identifiers. It will establish a central agency that will be able to collate data because, ultimately, AI is data driven, as I know the Deputy knows very well.
There have already been one or two projects, despite the lack of national infrastructure we need. Last December, the Mater announced it was the first hospital to use AI for patient scans. I might give the Deputy more detail on that in my response.
I note the Minister referred to the Mater. I understand it was one of the first, and probably the first, to integrate AI in its radiology department. I understand it allowed the hospital the opportunity to go through large numbers of results very quickly, and possibly not just to identify different issues but to get people additional tests while they were still on-site. There are real possibilities in that regard. If there are other examples of where AI is being used in a healthcare setting in Ireland, I would be very interested in the different ways that it can be used.
The Minister referred to the way that AI can also be used to look at large numbers of people, preparing the various records and having electronic records for that. One of the big challenges is that security and confidentiality will be needed, and buy-in and consent from patients for their data to be used in that manner. Has that been examined? Is there an update on that?
It is such a good question. There are many parts to this. One is general cybersecurity. It can be imagined, following the cyber breach we had, there has been very significant investment in and strengthening of cybersecurity across our health services.
There are many subtleties to AI, for example, ethical AI. We would be handing over some form of advisory capacity, decision-making capacity, or whatever it might be, to algorithms. There has to be appropriate oversight of that. Not only that, there has to be appropriate input to the design of that in terms of ethical AI guidelines. A framework for ethical AI guidelines was brought to the Government very recently. It is something we have to approach carefully.
The Mater project was clinically led. It trialled the use of AI software and looked at 15,000 patient scans to assess the potential to speed up care in the emergency department. Early results indicate improved diagnostic accuracy, streamlined workflow and reduced time required to provide life-saving treatments, so initial indications are very positive.
I understand that AI can also be used to customise care plans for people, looking at how they may respond to different therapies and customising the treatment or medicine from there. Have such considerations been looked at in any settings in the country?
I return to the huge issue of consent and the whole ethical question. If we are looking at having a huge store of people's data and running AI over it to predict where different populations might be impacted by various conditions, we need buy-in from people, trust in AI and consent for the use of their data.
There is also the question on the having it safe and stored securely as well.
It is an ongoing process and it is a rapidly emerging field. There are areas we are looking at very closely such as cybersecurity, as we discussed, ethical AI and the guidelines on the use of that, patient consent, and full information on how information may be used. We can do population-based analyses with or without AI. Having that population-based information will be invaluable for public health responses and investment in public health. There is no doubt AI can enhance that. There also would be AI applications for individualised care around genome sequencing and matching potential treatments to potential conditions that people have. This area will need very clear consent and regulation. The health information Bill is very significant legislation and a lot of the regulatory framework for digital health and e-health will be contained within that.