An ethical approach to data will help unleash the potential of AI in the Irish public sector

More than half (54 per cent) of Ireland’s public sector bodies have implemented Artificial Intelligence (AI) solutions in their organisation, while more than 30 per cent view AI as highly important for qualifying decisions and assuring quality. These are among the key findings of the ‘Artificial Intelligence in the Public Sector - How Public Organisations Benefit from AI’ report recently produced by Microsoft and EY. 

This positive attitude to AI and technology generally is very welcome, according to Frank O’Donnell, public sector lead with Microsoft Ireland. “AI is built on a foundation of increased availability of data, greatly enhanced computing power and very sophisticated algorithms,” he says.

“The Irish public sector is embracing this technology to help increase efficiencies and future proof multiple areas of the public service. With an ethical approach to data, AI can help the public sector to reimagine healthcare, transform the transport system, and improve the delivery of essential citizen services.”

Among the specific benefits he points to are better diagnostic services as a result of access to more insights about the patient and past patients.

“One of the areas where the power of AI is really relevant is education,” he adds. “It can support the education system in dealing with increasing numbers. And as we move to the hybrid blended learning model, AI can help understand how individuals best learn and help education providers fine tune delivery to get better learning outcomes.”

The security and integrity of citizen data must be protected at all times, O’Donnell points out. “There must always be safeguards in place. All use of citizen data must be very transparent from the off. That must be the first thing considered. The question for key public sector organisations is not what you can do with data and AI, it’s really around what you should do.”

According to John Ward, emerging technology leader at EY Ireland who conducted the research, “When it comes to ‘getting AI right,’ there is no secret to success. Rather, the successful integration of AI relies on investment in the fundamentals. Data and technology are more obvious and widely-acknowledged. However, culture, talent and ethics are equally important. With the right approach giving due attention and investment to each of these areas, organisations have the best chance of success.”

The National Transport Authority uses AI to keep Dublin moving

One public sector body which has successfully embraced AI is the National Transport Authority (NTA). “At the tail end of last year, we started to look at how we use AI and machine learning,” says Mark Stopes, head of business intelligence and data analytics with the NTA. “One of the use cases is how people use public transport. We can use data to analyse the origination of groups of passenger journeys, but the possible destination of these journeys has always been a big gap in our knowledge of how passengers use the public transport network. We have started to use machine learning to estimate where groups of journeys start and end.”

One of the big wins from that was much better insights into bus loadings at different times of the day. This could help to adjust schedules or even warn people in advance if a bus is full. 

“This is helping our understanding of why buses are full, where and when,” says Stopes.

This has assisted with the NTA’s response to Covid-19. “That’s been very interesting,” Stopes adds. “Covid-19 came before we got a chance to make use of it. Dublin Bus saw passenger numbers  dropping dramatically to about 10 per cent of normal. What’s happened is we’ve been able to use these new insights in a slightly different way. We are using the same basic tools to help Dublin Bus manage the capacity restrictions during Covid-19. That’s been a big success over the past six months.”

To promote sustainable transport, we have to convince more people to choose an end-to-end experience that works for them

He sees huge potential for the future of AI. “We will be able to use it to analyse things like queuing at stations and at stops. We will be able to use predictive analytics to inform the public about crowded buses. We could use telemetry to understand maintenance plans and watch out for the tell-tale signs of maintenance requirements for brakes and engines and so on. This will help reduce the amount of time vehicles spend off the roads and rails. The NTA recognise that to deliver features like these require additional investment and concept proving, however the benefits to our customers are clear and the effort worth the investment.”

Ultimately, it could help with the country’s climate change goals. “In order to promote sustainable transport, we have to convince more people to choose an end-to-end experience that works for them. When people get into their car there is a degree of certainty around departure and arrival times. AI will help to bridge the gap with public transport by helping to understand and predict what services to use, advise on when to leave the house, the reliability of the service they are using, if there are seats, and so on. People can make their own decisions based on that. Used in that way, AI can help make society more carbon efficient and ultimately reduce our carbon footprint.”

Microsoft has been working with a number of public sector bodies on AI adoption. “We work with our customers to build and evolve solutions to their needs,” says O’Donnell. “We are continually bringing AI into our own solutions like Office and Teams to make them better. We can share that experience with our customers. We also have lots of partners with expertise in particular areas. They can show our customers the art of the possible and share with them what’s happening in comparable jurisdictions.”

Investing in data sets that reflect the diversity that exists in our society is vital

Trust is paramount for the successful adoption of AI, he adds. “We have a very strong ethical approach. For any entity – whether it’s a local government agency, a hospital or an educational institution – building and maintaining trust is absolutely critical to ensure employees and the citizens they serve feel comfortable working with AI. Establishing clear guidance and transparent processes is an important step. To help address this, Microsoft has developed guidance and training on responsible AI design and use which we share with our customers. We have also inputted into the Government’s AI National Strategy consultation and highlighted the absolute necessity of taking an ethical approach to AI.”

Mark Stopes explains that the NTA is not interested in individuals’ personal data. “We only ever use data at a cohort level,” he says. “The thing about AI and the data we collect is that we are not interested in the individual. We are looking at how people move in masses. We have very strong data protection safeguards in place around the things we do. First and foremost, we never look at individuals. AI and machine learning don’t work at the individual level in any case.”

O’Donnell believes that Covid-19 will accelerate the adoption of AI in the public service. “Our public service customers who were operating at the nexus of the crisis in mid-March came directly to us and now we are working in partnership to deliver the best solutions for citizens. This may prove to be an inflection point for technology in the public service. The public sector has delivered in incredibly difficult circumstances. Public bodies put in systems in weeks that would normally take months or years. Our public servants should take great pride in that achievement and how it has impacted our citizens,” he says.

To download the ‘Artificial Intelligence in the Public Sector - How Public Organisations Benefit from AI’ report recently produced by Microsoft and EY, click here.