7 minute read 28 Feb 2019
Businessman commuter riding bicycle urban bike lane

How AI can be a force for good in government

By George Atalla

EY Global Government & Public Sector Leader

Working with governments to address complex issues and build a better working world.

7 minute read 28 Feb 2019

AI has the potential to overcome the biggest challenges governments face and dramatically improve life for citizens.

Around the world, companies are using artificial intelligence to automate mundane tasks, make better decisions and improve the customer experience. But AI isn’t the unique preserve of the private sector. Governments also recognize the potential of these technologies to transform the way they organize themselves and run their processes. And they’re waking up to the fact that citizens expect the same revolution in services that they’ve experienced from the private sector.

That means governments having a single view of citizens’ data, and sharing it in a relevant way across departments while protecting their privacy. It also means using that data to create new services, anticipate what citizens will need next and take action to prevent crises.

“AI represents an opportunity for countries of all sizes,” says Keith Strier, EY Global and EY Americas Consulting Leader for AI. “Larger economies may have an early lead, but small countries can compete, if not 'punch above their weight' in targeted or niche AI applications, creating a new basis for global competition delinked from traditional measures of national power and strength.”           

AI offers governments two big opportunities that don’t apply to the private sector:

  1. It allows them to structure and analyze the huge amount of data they hold on citizens – and use it for social good. This means they can quantify and reduce inequalities in outcomes as well as opportunities. They can also share the data with third parties, who can create apps or services that improve life for citizens, while making sure those parties keep the data private.
  2. It gives them a unique chance to drive how citizens use and benefit from these technologies. That’s because governments are also responsible for role-modeling the ethical use of AI, regulating how companies apply it and educating citizens to be ready for its challenges.
Small countries can compete, if not 'punch above their weight' in targeted or niche AI applications.
Keith Strier
EY Global and EY Americas Consulting Leader for AI

But by deploying AI, governments expose themselves to the same risks as companies –  such as building bias into algorithms. And because of their regulatory role, it only takes a big data breach, such as the recent compromise of personally identifiable information of 30,000 US Department of Defense personnel, to damage trust in government irreparably.

This complex picture is why we recommend that governments use a “trusted AI” framework. This will make sure they not only consider how an AI-based system performs, but also identify and mitigate the risks inherent in every stage of the solution. For example, they can make it clear to citizens that they’re dealing with an algorithm as a service, so they can opt out or transfer to a human.

What exactly is AI and how can it help governments?  

The growing expectations of citizens is just one of the challenges facing governments today. Rapid urbanization, an aging population and complex socio-economic problems are putting public services under strain. And with low economic growth keeping budgets down, governments need to find solutions that are both efficient and sustainable.

AI is a set of technologies and capabilities that can help governments to solve these challenges. It does so by supplementing certain human competencies or, in some cases, replacing them.

It consists of three main areas:

  1. Sensing. AI can augment or replace human sensory capabilities, speeding up simple tasks such as visual detection. For example, AI software can automatically analyze street and traffic cameras in real time. So governments can make the best use of public transport, reduce pollution and manage the flow of traffic.
  2. Thinking. AI and related technologies, such as machine learning, deep learning and natural language processing, can analyze and process large volumes of data much faster than humans, and in some cases, more effectively. Some governments are already using these technologies to help teachers fill the gaps in teaching and learning – for example, by carrying out admin tasks and tailoring learning to different pupils’ needs.
  3. Acting. AI and related technologies such as intelligent automation (think virtual assistants or chatbots) can take simple decision-making tasks off humans. This frees up time for front-line workers to focus on activities that improve services and the citizen experience. During the Winter Olympics in South Korea, for example, several humanoid robots, equipped with AI-powered translation software, were used to provide information to visitors and athletes.

What will AI mean for jobs?

Applications such as these will help to build citizens’ trust in governments and overcome any public perceptions that they’re difficult to deal with. But to maintain that trust, governments will need to make clear to both employees and citizens that AI won’t spell the death of all jobs. In fact, in its recent repor (pdf), the World Economic Forum noted that AI technologies could displace 75 million mundane or repetitive jobs. But they could also create 133 million new jobs, which would be more skilled or creative.

For example, AI technologies can recognize, understand and draw insights from vast amounts of unstructured data. As a result, government officials, particularly in planning, HR, customer service and procurement, will effectively become the managers of AI models – monitoring and adjusting them when needed. They’ll also be able to use the insights they gain to make more informed decisions and allocate resources better.

So while AI may remove the need for routine or manual labor, overall, it’ll augment human intelligence – not replace it.

AI in action

It’s our view that governments that use AI wisely can overcome current challenges and transform life for citizens. Over a series of articles, we’ll look at six areas in which they can do this. We’ll also cover the need for governments to regulate AI, and how they can roll out AI-based systems effectively.

We’ve listed the six areas below and provided a real-life example for each.

  1. Improving public finance management
    The UK’s Department for Work and Pensions has tested the use of algorithms to identify and crack down on criminal gangs defrauding the benefits system. The algorithms have been able to detect cloned identities among billions of items of data. By rolling them out across the whole benefits system, officials hope to prevent the loss of millions of pounds..

  2. Building the cities of the future
    Singapore is working hard to get self-driving taxis on the road as part of its Smart Nation initiative. It’s even built a mini town as a test park for its driverless buses. The aim is to lower people’s dependence on cars, reduce greenhouse gas emissions and make it quicker and easier for citizens to get around.

  3. Making life safer for citizens
    The Kanagawa Prefectural Police in Japan plan to trial predictive policing before the 2020 Tokyo Olympics. Their AI-based system will analyze big data to see if the same perpetrator is behind several crimes. It’ll also predict where crimes are likely to happen and when. By acting on its insights, officials will reduce crime and make Tokyo safer.

  4. Improving national defense and security
    Researchers at Massachusetts Institute of Technology in the US have worked with a tech start-up to develop an AI system that can detect 85% of cyber attacks. The system will help officials to prevent security leaks by making decisions early. In the long term, they’ll also be able to improve their security systems and prevention strategies.

  5. Giving citizens a better experience
    Local governments in Denmark are using AI to analyze posts by citizens on their social networks and identify the biggest issues they face. This allows them to respond to issues proactively and use predictive analytics to address problems before they emerge.

  6. Helping vulnerable citizens such as children, the homeless and substance abusers
    We’ve worked with a London borough to pilot an AI-based system to identify children in need of safeguarding (pdf). The system had an 85% success rate, and in just seven months, it identified 1,700 families that would benefit from targeted services.

Are you ready to adopt AI?

Before your organization and citizens can start reaping the benefits of AI, you need to answer some fundamental questions. These include:

  • Do you store and manage your data in the cloud?
  • Can you integrate data from diverse agencies?
  • Have you created relevant roles in your organization, such as chief data or digital officer?
  • Do your employees understand what AI technologies you’re introducing, why, and how the change will affect them?

Answering these questions will help your organization have a clearer understanding of its current capabilities and the type of work needed to move ahead.

Summary

Governments can use AI to become more efficient and effective, as well as offer citizens better, more personalized services. But it’s also government’s responsibility to role-model AI’s use, regulate how companies apply it and educate citizens to be ready for its challenges. And crucially, government needs to identify and manage the risks that AI poses.

About this article

By George Atalla

EY Global Government & Public Sector Leader

Working with governments to address complex issues and build a better working world.