Official Report: Minutes of Evidence

The Chairperson (Ms Bradshaw): I welcome to the meeting Professor Helen McCarthy, who is the Chief Scientific and Technology Adviser.

Professor Helen McCarthy (The Executive Office): Thank you very much, Chair.

The Chairperson (Ms Bradshaw): Thank you very much for the paper that you provided in advance. If you would like to, go ahead and make some introductory remarks.

Professor McCarthy: Sure. No problem. Thank you for giving me the time to update you on my work.

As you can see from the evidence paper, we have a number of different work streams. You will be more aware than I am of what was in the Programme for Government (PFG) about setting up an Office of Science and Technology and the Office of AI and Digital. We have really been focused on building the Office of Science and Technology. There are quite a few work streams underneath that, but I will give you a bit of context around it.

One of the first things that we did was to put a science champion in each Department, because we felt that that was really important. Many people would ask, "What is the big deal about research and innovation?". Essentially, by bringing in more R&D, doing more research and innovation and having more solutions, we create better jobs and our society becomes more skilled. It is really about elevating our aspiration to put R&D across all disciplines. That has very much been a mantra in what I am about.

The departmental science champions were responsible for putting forward areas of research interest (ARIs). While research interests have been going on across Departments, we perhaps have not been the best at communicating those to academia and the wider sectors. I am pleased to say that we have 74 areas of research interest across all Departments. We plan to put those on the ari.org website, a national UK database, and we are setting up our own NI Science website. That is about showcasing what we have in Northern Ireland and ensuring that people can access the departmental science champions and interact with them. If they put up what those research areas are, that is an opening for a conversation with the broader science community. Areas of research interest are one thing.

I have just come from an event for the International Day of Women and Girls in Science. I have set up the women in science, technology, engineering, arts and mathematics (STEAM) organisation in collaboration with the DAERA, the Department for the Economy and the Department of Education. They have co-sponsored that over the past 18 months. The idea behind it is that the arts and humanities are critical as a discipline in the science, technology, engineering and maths (STEM) subjects. We need to have creativity, and we need to bring our civil servants with us on that journey. We have been focused on educating that cohort of women about technology, innovation, AI, our ecosystem and what it means to them, as well as how it affects them. We are starting to open those conversations. We have a Teams channel for women in STEAM. They have made connections and joined the dots. While our Departments have their own remit, a lot of the technologies that we face span them, so we need to start collaborating.

We have also developed the first combined research and innovation strategy. We had a call for evidence last year, and we used AI to analyse that evidence with the support of PwC. We put together the strategy, which has gone around the departmental science champions and the permanent secretaries and is now with the Executive Office to go out for consultation. A lot of the actions in the research and innovation strategy, as you would imagine, affect further and higher education. A lot of those sit within the remit of the Department for the Economy. I have been working closely with colleagues there to make sure that we can implement them.

One of the exciting things that have come out of that is the idea of an innovation panel. I can talk to that a bit more, and you might have some questions about it. It is the idea that DFE, the councils, the universities, business organisations and cluster leads will all make a decision about what the country needs going forward in respect of research and innovation and the structures that we need to ensure that we are not reacting to the calls from the Department for Science, Innovation and Technology (DSIT) or from whichever body it happens to be and that we are proactively engaging in seeing what opportunities we need.

One of the examples that we are using the innovation board for is the local innovation partnerships fund, which is Sir Patrick Vallance's scheme. Some £30 million was allocated to Glasgow and Manchester, and we said, "What about Northern Ireland?". We have managed to bring that fund in, and we are working collectively around all that.

NI Science is an interesting one. Everybody in the room is much more experienced than I am, so I do not need to tell you that we are an SME-driven economy. We have a lot of really exciting young companies, but they do not always have access to the equipment that they need because it is really expensive. The idea with NI Science is that it will be a country-wide core technology unit where we put out information on all the public-sector equipment that we have that companies can come and use. That will help to form collaborations and innovations. It showcases the science and the test bed that we have. It also showcases our departmental science champions. We also have government AI champions.

Finally, the AI strategy is probably the one that everybody is interested in. We had 13 meetings across 95 stakeholders in the spring/summer of last year, and then we narrowed it down to an external reference panel. We have drafted the strategy. There are 72 actions, and we can discuss those in more detail. On the basis of that, it went to the permanent secretaries just before Christmas. The feedback deadline was the end of January. We have received all the feedback, and we are now preparing to go out for consultation.

That is it in a nutshell. I am happy to have a conversation with anybody about any of the work.

The Chairperson (Ms Bradshaw): Brilliant. Thank you so much. That was my first question, so I can tick that off the list.

I will pick up on the concerns that some people have around AI and the potential to take over jobs. There was an announcement yesterday about some job lay-offs. Is it your responsibility to undertake work with the universities and industry to future-proof jobs and ensure that we have a workforce that keeps pace with the moving technologies?

Professor McCarthy: That is a good question. The strategy is about AI in the public sector. We need to be clear that the essence of the strategy for the public sector is positive use, positive inclusion and positive outcomes. We are saying that artificial intelligence is never a replacement for human intelligence. That is really important, and we need to understand that: AI does not replace humans. Humans put the input in: input in the quality of the data, input in the checking as you go along, so that you do not have any bias, and input in how much you want your AI system to do. We are not advocating that any AI system that is used in the public sector should make human decisions, because we do not think that that is appropriate. However, it can streamline processes, which can free up people's time so that they can reskill and move to other areas that they might be interested in. In TEO, for example, over 200 people have undergone training in Microsoft Copilot, and a further 100 people are training in M365 Copilot. Again, it is about using that to streamline our systems and processes. That is really important, but that is different from business AI.

You will know that the Department for the Economy and both universities have set up the Artificial Intelligence Collaboration Centre (AICC). Ulster University is the centre's lead partner. The main aim of the centre, which is driven by David Crozier, is to create a test bed for SMEs and help them to embed AI in their systems. They have different rules and regulations, so they can have their sandboxes and test beds to try things out, but, in the strategy for the public sector, I am clear that, unless it meets the ethical standards that we require in the public sector, it should not be adopted in the public sector.

I am advocating strongly for a public registry of all the AI projects that we have, as well as a citizens' forum in which we can talk openly about what is going on in AI. Having AI champions in each Department is about starting to talk about AI, drive the changes and register some of the ongoing projects. Some exciting things are going on, but we need to make sure that every Department has a human oversight team that oversees all the projects. What I mean by that is that you do not have one person responsible for one AI project and another for another; you have to have some sort of oversight team that checks that the data that is going in is of good quality, that you are not getting any drift and that your outputs are really good.

The Chairperson (Ms Bradshaw): I represent South Belfast, which has Queen's University at its centre. It has carried out amazing research over many years to do with the public services and public policy. What will your role and the role of the champions be in better accessing some of the theses and dissertations on public policy so that we can be better informed? I am conscious that a lot of work has gone on. How will you be able to distil that through AI?

Professor McCarthy: We have 160 academics across both universities. I believe that there have been over 500 outputs — publications etc — in this space. As part of the external reference panel that we had for constructing the AI strategy, we had representation from Queen's University, Ulster University and the regional city deals. It is very much a co-creation, and that is the right thing to do. The innovation panel that we put together comprises government, industry and academia, all working together on that.

The Chairperson (Ms Bradshaw): I am not sure whether I explained that properly. I am talking about some of the research into, for example, adolescent mental health. I am talking about different areas and not purely about AI. Is that what your answer was about: the various areas of public policy?

Professor McCarthy: Do you mean more general public policy?

The Chairperson (Ms Bradshaw): Yes.

Professor McCarthy: The idea behind having departmental science champions is to open that up. I will give you an example. Obviously, I am coming more from a science background than a public policy background. An all-Ireland cancer data conference took place the other week. It was a two-day event set up by Professor Mark Lawler from Queen's University. That was great to see, because we had the permanent secretary from the Department of Health, as well as the department's Chief Digital Officer, alongside the Irish Government's science adviser, Aoife MacLysaght. We have monthly meetings to make sure we are joined up on that. We talked about how, culturally, government has taken a long time to open up to the academic world. Those barriers are being broken down. I do not know whether that answers your question about public policy.

The Chairperson (Ms Bradshaw): It is reassuring that people are getting in the room and coming together.

Professor McCarthy: For example, I liaise regularly with the pro-vice-chancellors in Ulster and Queen's universities about what is going on. For example, I have a meeting tomorrow with the pro-vice-chancellor of engineering and physical sciences in Queen's who wants to update me about what is going on and also, "Can you open any doors?" or "I can't get any contacts into DFI" or wherever, and we try to make those happen.

The Chairperson (Ms Bradshaw): The Infrastructure 2050 draft investment strategy is a responsibility of the Executive Office. It is really about prioritising investment over the next 24 years just to ensure that we keep pace. To what degree does the lack of such an overarching investment strategy in place impede you, or does it allow you to move at pace with regard to your work?

Professor McCarthy: The First Minister and deputy First Minister have been clear that they wanted a research and innovation strategy and then an AI strategy, so it has not really prevented us from doing the work —

The Chairperson (Ms Bradshaw): Your core work, OK.

Professor McCarthy: — so far, so that has been good. Now that we have a good road map for implementing the research and innovation strategy, we need to build our AI team. I think that that is recognised. We need to build that skill set and experience not just in TEO but across government.

One of the most important things that we have, which is critical in a large organisation, is that, culturally, people are saying, "We need to change". If you have people saying, "No, we're all fine. We don't need to change", it is difficult to implement change. Whereas, when you have people saying, "Yeah, I need to understand technology. I want to understand. I want to be trained", it is much easier to drive that change. People may think that it is not happening fast enough, but compared with what I have seen, there are a lot of really good projects on the go at the moment that should drive change.

Mr Dickson: Thank you for bringing insight into this. It is an exciting area of work, but there are lots of issues that, I hope, you are finding time to drill down into. I am interested that you have appointed or sourced science champions in each Department. Although we have champions for just about everything in Departments at the moment, there is not much evidence of any actual output from them. We have more champions in Departments than some people have had hot dinners. Anyway, let us hope that science champions will champion science in Departments.

I want to drill down into, for example, the Department of Health, because health has a broad range of science research and trials that are ongoing. I am not suggesting that it is the biggest, but, for our universities, it probably is the biggest area for research and trials that can take place in Northern Ireland. Things happen at a basic level; research is done; the trials are started; and, indeed, new businesses will start out of that as well. The way in which, for example, cancer trials are carried out in Northern Ireland — there was a conference last week — is labyrinthine. Each trust has to approve its own research from its staff or others. It all comes forward to the Public Health Agency (PHA), which then filters those projects, and they hit a brick wall because there are only one and a half members of staff there.

Every research project in health in Northern Ireland, including student research in the medical schools, comes into an in-tray. It is not prioritised; it is simply taken in date order to allow those projects to proceed. Once a project gets approval to proceed, it might often be part of a national project, because the universities or the health professional undertaking it will not be doing it on their own; they will be doing it in collaboration with colleagues across the United Kingdom, often across the Republic of Ireland and, of course, across Europe and, indeed, internationally. However, there are massive barriers to those things actually starting.

What I want to know — as the science champion, you have a great job to do — is what are you doing to resolve the detailed problems that create barriers and cause people's health not to be able to be progressed in Northern Ireland?

Professor McCarthy: There are a number of angles to what you said, and you are absolutely right. The first is that there is a Chief Scientific Adviser (CSA) in Health who has responsibility for that, and there will be a new CSA in Health when the current CSA finishes on 13 March. As you know, there is a new permanent secretary in Health, Mike Farrar. He has convened an innovation group, and I, along with academics from both universities, the Chief Pharmaceutical Officer, the Chief Digital Officer and the chief executive of the Health Innovation Research Alliance Northern Ireland (HIRANI), sit on that, and everything that you have said has come up in conversation.

Mr Dickson: Good.

Professor Helen McCarthy: We know that there are things that need to be done.

Mr Dickson: At the moment, the system is, at best, not working, and, probably, it is broken. It needs a radical rethink if we are to get ahead of a lot of the problems.

Professor McCarthy: Agreed. The Department of Health is absolutely all over that. I have seen a real change even in the 18 months since I have been in post. I am at the table with everybody. I was at a round-table conference meeting yesterday afternoon in Queen's where they were talking about academic clinical partnerships. Everybody is talking about that, and everybody knows that there are actions that need to be taken. You also know that one of the regional city deals is the iREACH clinical trial centre. Now is a good time to start to ask how we do things differently.

Mr Dickson: Do we have a timeline and a plan for delivery of that improvement across the Department of Health and out into the world of academia and the wider world of research?

Professor McCarthy: I do not know the answer to that, but I will take it as an action to find out from Health what the timeline on that is.

Ms Murphy: Thanks, Helen, for coming in to brief us. The Chair and Deputy Chair have touched on some of the questions that I wanted to ask, so I will not be too long.

Helen, can you expand on what some of the ARIs are?

Professor McCarthy: Obviously, each Minister in each Department has their priorities, and there will be research that underpins some of those. One of the Department for the Economy's key focuses is on how to extend economic development outside of Belfast to wider places, so it is interested in ARIs for what research it needs to do to figure out how to do that a bit better. Similarly, in TEO, you will know that there are priorities on, for example, ending violence against women and girls. It is about making the Departments realise that they need to do research to develop what they want to do to find the answers. Sometimes, we can all be guilty of thinking that we know the answer to everything, so it is about extending that Department out to all of the academics who have written the great papers on policy or have the technologies required. What is interesting is that, of the 74 ARIs that have been put together, over 70% of them have an AI/digital focus to them. Departments are now actively saying — it will mainly be the universities that benefit — that they need help in that area. That is really encouraging.

Ms Murphy: I find all of this stuff absolutely fascinating, Helen. It is the issue of trying to marry how government formulates policy with the academia end of research and development. The women in STEAM stuff is absolutely brilliant as well. You mentioned the all-Ireland cancer data conference: is there potential for further cross-border collaboration?

Professor McCarthy: There is huge potential. It takes place in quite a few areas. There is the all-Ireland cancer data network. Bringing government officials into that has definitely opened up opportunities and possibilities, which is great to see. As I said, a big part of my role is connectivity and making sure that Northern Ireland is represented, so having a monthly meeting with Aoife?McLysaght, who is the Irish CSA, is really important to see how we can help in that research space. I also have monthly meetings with the Scottish and Welsh CSAs to make sure that, when there are wide calls, we have our ducks in a row and we know what our strengths are and where we are going. Another example of where there is particular opportunity North/South is around data and data storage. Obviously, we have Encompass and all the healthcare data from that. That is a massive asset. Those discussions are ongoing, and there are huge opportunities to bring everybody together, but the trick is that we have to get people to understand that a smaller slice of a big pie is more important than a big slice of a small pie, because the pie will be much bigger. The research and innovation strategy is called 'Integrate to Innovate', because you cannot do it on your own. You need to have government, industry and academia working hand in glove. That is really important.

Ms Murphy: Helen, you mentioned important ideas: integration and innovation. That will be important as we go forward with the work. Thank you very much.

Mr Gaston: The way in which everybody does business is changing because of AI, and, if you do not get on board with that, you will be missing a trick when it comes to some of the benefits that it can bring. I read in today's paper that the Department of Health and the Department for the Economy are investigating ways to use AI to answer Assembly questions for written answer (AQWs) from Members. It is my understanding that AI uses data that is already available in the public domain. I get that safeguards will be put in place, but, if that idea goes ahead, how confident are you in how fulsome the answers would be? My initial concern is about data breaches as a result of depending too much on AI. We saw human error in the PSNI data breach and how much that cost. I am concerned that, by trying to upload a lot of information for AI to capture to respond to AQWs, we will risk repeating data breaches in the future.

Professor McCarthy: That is a good question. As you know, each Department will have a data protection impact assessment (DPIA) process. We are saying that AI has to be included in the Data Protection Act 2018 and has to be embedded in it. The question is not about the use of it; it is about where they draw the information to use from. At the moment, the information on Microsoft Copilot is all closed within the system. Using Microsoft Copilot in the Civil Service is not the same as using it outside the Civil Service. It is all closed, and it can basically be guaranteed that it will not do hallucinations or make information up. The main concern in all of this is that you do not want hallucinations to happen. Think of it this way: using an AI tool to pull out the evidence that you need to help inform your answer is OK, but, when it comes to drafting the answer, we need to make sure that we have that human oversight and that we make that final decision about what is in the answer. Using it to gather the evidence is really good because that saves time, but you have to make sure that the source is robust and is closed within the system. That is the main thing.

Mr Gaston: Obviously, MLAs and office staff have access to Copilot, but, when the security team came round, it recommended that that is the only AI tool that we use.

Professor McCarthy: Yes.

Mr Gaston: Are MLAs part of that closed system?

Professor McCarthy: You are, yes. Absolutely.

Mr Gaston: If AI were to start to pick up information, would it pick up stuff outside of Departments and draw from that? How does it distinguish that the information for, say, the Department of Health is coming from only the Department of Health?

Professor McCarthy: I do not know the answer to that question. I would have to talk to DOF about how it has been set up. The theory is that you should be able to draw your information from your Department. I do not know whether it has a cross-departmental function. I need to find that out.

Mr Gaston: I am very open to the idea. That is the way that we need to go because it is more efficient. Your strapline was:

I agree with working smarter to get the information that we need. We have had examples of Ministers using copy and paste. The Minister for Infrastructure responded to AQWs with, "Blame the Brits for no money". If we can get AI to provide Members with the answers to questions, that would certainly be a step forward, as long as it is done in a safe way that protects people's data.

Professor McCarthy: The quality of data that goes in is important. The data information officers in the Departments have a key role to play in making sure that what they input is fit for purpose, and that will be key. People get obsessed with AI and its uses, but it is about making sure that the data going into the system is correct. The data architects and data analysts are really important, and that skill set is required across every Department.

Mr Gaston: Thank you.

The Chairperson (Ms Bradshaw): Thank you so much. It was good to have you at the Committee. If you want to update the Committee on any reports, please send them through.

Professor McCarthy: There would be questions to answer. [Inaudible.]

The Chairperson (Ms Bradshaw): Thank you, Helen.