Four ways for AI to change government

By Apala Bhattacharya

The effects of AI on public services? They become more efficient, personalised, predictive, and bottom-up.

Will AI become so intelligent that we don’t need government any more? That’s the suggestion from Taiwan’s digital chief, who notes that “We can replace the entire government by AI in 500 years”.

That seems unlikely to affect many GovInsider readers - unless modern medicine dramatically increases the average human life expectancy.

But already we are seeing this technology improve public service delivery. And far from making government become colder and more mechanised, this ancient institution is getting closer to citizens than ever before.

1. More efficient

Image: KennislandCC BY-SA 2.0

There is a prevalent notion that the public service are more often than not too slow to respond timely. They are slow on the uptake, and even slower on delivery.

At times this is in fact true, due to constrained budgets and resources governments might be unable to appropriately direct funds and human skills to the areas it is most required.

With the use of AI technology, governments are becoming increasingly more efficient both in the time it takes for them to respond and in the allocation of resources.

In the US, the army is using interactive virtual assistants which through deep learning helps answer any queries, reviews qualifications, and sorts candidates. These potential hires are then referred to human recruiters, thereby speeding up the process.

This also frees up public service officials to dedicate their time and effort to more complex problem-solving instead of sorting through thousands of applications.

Across the Atlantic, the Danish government is attempting to use AI technology to help process applications for grants and welfare payments. The technology will allow officials to make decisions faster and remove the need for a case worker to pore over the papers which might slow down the process and introduce biases into the process.

This digital grant application process will also include support for the elderly, as well as funding for low income families.

2. More personalised

Image: David B. GleasonCC BY-SA 2.0

The time that is saved when AI technology is utilised in public services can help public servants personalise government towards each citizen. In New Zealand, a predictive service called SmartStart provides personalised information from across all government agencies to expecting parents. The platform allows citizens to fill up various forms, and even apply for a birth certificate-all from their mobile phones. This leaves parents to enjoy quality time with their newborn, instead of running from one government agency to another.

A no-show predictive model applied across various public hospitals in Singapore help predict the no-show rate of patients using data analytics and artificial intelligence and thus personalises patient interactions. A reminder text is sent to patients with a high rate of no-shows. The programme can also check with patients if a reschedule is required.

The service allows patients to better interact with public healthcare and reduce time wasted on menial administrative tasks.

3. More predictive

Image: IAEA ImagebankCC BY-SA 2.0

The UK Home Office has funded several developers to create an artificial intelligence programme that takes down content uploaded from terrorist organisations, and learns these patterns to help prevent them from being uploaded in the first place. This makes terrorist propaganda harder to access, and reduces the radicalisation of individuals. Thus, it lessens the likelihood of lone wolf attacks. AI technology provides public service an edge over rapidly changing and developing security threats.

In Malaysia, the police force and the national ICT agency have teamed up to create a surveillance system in prisons that can detect and predict inmates’ aggressive behaviour as well as escape attempts. The system also takes note of the behaviour of police officers on duty to protect inmates from maltreatment. Once triggered, the system will alert officers on duty and direct them to the potential breach.

4. More bottom-up

Image: sese_87CC BY 2.0

In Las Vegas, the health department trialled a software that collated and analysed thousands of tweets to identify potential health inspection sites for restaurants. The program calculated the type and likelihood of health problems from words and phrases off of tweets, which was then flagged and connected to specific restaurants.

Health inspectors were then sent to these restaurants to conduct thorough checks. The trial led to a 6% increase in citations in comparison to when inspections were carried out at random, and allowed information from the civil society to guide the action required.

In Taiwan, legislators have used an AI-powered platform to bring together over 4,000 different stakeholders over the issue of ridesharing apps in the country.

Riders, taxi drivers as well as Uber drivers deliberated over the regulations Uber and other ridesharing apps should follow in Taiwan, while the platform kept track of and grouped people with similar viewpoints. After over a year of deliberations, Taiwan ratified seven new regulations agreed upon by all parties.

AI might indeed take over government in the next 500 years, but we do not have to wait that long to see its positive influences across the sector. It already is reducing inefficiencies in public service delivery and making government more accountable, approachable and accessible.