Governments are already using data and analytics in a number of ways to help them become better informed and provide superior services for their citizens. For both central and local governments, an increasing number of back end processing and citizen engagement opportunities are emerging for smart use of artificial intelligence and its many subfields.
The biggest area for potential quick wins will be the vast processing that occurs in various administration tasks. This includes improving awareness of patterns in data, to create new theses and models. Bringing together data from different areas and using algorithms that learn, can create new insights. These insights, in turn, can help public service delivery teams to create new models about how their citizens behave, and therefore improve the way in which they provide services. Insights might be about ‘known’ and ‘unknown’ issues, leading to very different ways of working.
Helping individuals to make better choices is another important area. Public services are ultimately all about their populations, and populations are made of individuals. Individuals making better choices—that is, both better for them and for the community or population—can therefore change the environment, one person at a time. Machine learning, coupled with good data visualization technology, can be used to show citizens the consequences of their decisions quickly and simply, and help them improve their decision-making for everyone’s good. It is a potentially powerful augmentation to policy development and/or improvement.
Health and space for wellbeing
Healthy eating may not sound like an initiative that requires AI support. But a healthy population has fewer health problems, resulting in less demand for healthcare and nursing infrastructure. And preparing for the demand of services of the aging population, this is not an insignificant matter. The city of Amsterdam, for example, has used data from grocery stores about vegetable sales to evaluate a city campaign to encourage healthy eating especially among children.
Supporting better planning for land use around urban areas is an emerging use case. Demand for housing is growing exponentially, according to many government models. Better analytics about demographics, income levels, businesses and general land use within an area can help planning officers to ensure that urban brownfield land is used effectively to meet the needs of the population in the area, rather than just to suit the developers’ ideas.
Improving the use of energy by predicting spikes in demand is one example. Artificial intelligence systems can learn from experience, which means that they can analyze past patterns in data, and use them to predict the future, and therefore manage better. This is particularly useful in managing power grids, because the system can learn when spikes in demand are more likely to occur, and enable better use of the power at times of lower demand.
Variable energy pricing could be a way to a more careful use of energy by individuals. Artificial intelligence systems can be used to propose variable pricing models to encourage use at low-demand times. For example, the UK’s Economy 7 tariff is a good example of a tariff that increased night-time use of energy for heating.
Better transportation management
Control parking by variable pricing on parking meters is already planned. Boston is planning to raise the parking prices for the first time in more than five years. But it is not just planning a blanket hike. Instead, they will use the changes to test how parking prices affect occupancy of parking bays, and the way that people use cars. San Francisco also plans to implement demand-driven pricing as a way to control congestion.
Creating smart parking systems to direct drivers to spaces, and reduce congestion is an interesting possiblitity. One problem in cities is congestion created by drivers driving around searching for a parking space. Deutsche Telekom is trying to address this via smart parking systems which will alert drivers about empty spaces, and direct them straight there. Systems are being tested in Dubrovnik in Croatia, Bucharest in Romania, and Pisa in Italy.
Intelligent traffic management helps to reduce congestion and air pollution. Sensors are not just useful for parking meters: pollution sensors on traffic lights and at bus stops can also help identify areas of congestion, and direct traffic away from those. Providing real-time information about hold-ups and traffic jams to drivers, together with suggestions for alternative routes, helps traffic to avoid those areas. This, in turn, reduces congestion, emissions, and air pollution in busy areas.
Vehicle-to-vehicle communications offer increased potential to avoid congestion. Smart traffic management can also involve direct vehicle-to-vehicle communications. Vehicles including cars and buses can be fitted with RFID tags that can communicate with each other. Information about traffic jams can therefore be passed on ‘from the horse’s mouth’, almost literally, helping drivers to avoid problem areas.
Tell us about your public service use case
We see many more exciting public services use cases ahead for the spectrum of artificial intelligence subfields. Will you join us on Twitter for a #saschat at 15hrs CET on Friday 24th March, to share your ideas and discuss drivers and inhibitors? We will use the following prompts to steer the conversation a little:
Q1: What re your favourite examples of AI in public services?
Q2: Who do you see as stakeholders driving better analytics in government?
Q3: What re the challenges facing public service teams pursuing better exploit of AI?
Q4: How should success of AI in government be measured?
Q5: How can citizens be supported to get more out of AI enriched services?