In the Easter tornadoes in Mississippi last year, emergency teams faced two competing disasters. As well as searching for survivors and spotting which structure might fall next, overworked Federal Emergency Management Agency workers also faced the threat of Covid-19. 

Putting boots on the ground could simply exchange the disaster of the hurricane for the catastrophe of spreading the virus. What followed was a key example of how artificial intelligence and data can solve a crisis by sharing skills and standards to create a smarter state, as well as empowering individuals for the future.

GeoX, a connected data and AI company, rapidly micro-analyses the structural integrity of buildings after natural catastrophes. Through satellite pictures and digital mapping, they were able to illuminate which structures had collapsed, which were damaged and which were at risk of falling, all without breaching the rights of the individual.

Looking ahead, GeoX will unite the public and private sector, mapping every building and structure in the US and Australia for risk of collapse, arming individuals and the state with the information needed to make homes safe, says Guy Attar, the company’s co-founder and chief business officer. “The future will be advanced analytics, creating accurate risk scores for all kinds of different perils across global geographies,” he says. 

This is one of a growing number of AI applications that are saving and enriching lives. With greater collaboration globally and between public and private sectors, states will become smarter. The opening and sharing of data will supercharge governments and local authorities, but we aren’t there yet. There are challenges to overcome.

Harnessing the power of AI

AI’s ability to find a needle in a haystack isn’t just useful in disaster zones – it can also prevent them. Ontario Power Generation is one of the largest low-cost, clean power generators in North America. It relies on AVEVA’s AI-driven predictive analytics to improve overall performance. 

However, it also spotted a potentially catastrophic failure on a multimillion-dollar steam turbine, says Jim Chappell, AVEVA’s global head of AI. Engineers assessed the problem and found nothing, but the AI persisted in warning of a brewing crisis. “By itself, the drop in vibration was not something that would have been noticeable, but had this turbine issue not been caught, the blades would have eventually been destroyed, resulting in catastrophic damage to the turbine and costly, extended downtime,” says Chappell.

As well as saving an estimated $34m (about £25m) in maintenance costs in this instance, similar “plants of the future” can dramatically impact climate change, reducing the 35 per cent of carbon emissions that currently come from manufacturing, according to the UN.

Additionally, AVEVA digital twinning is creating “connected workers” armed with cloud-based AI technology. This doesn’t just spot problems; it produces deep, future insights. For example, it’s helped to reactively and predictively guide ambulances through busy traffic in India, as well as spotting pipe leaks in the Niger Delta, warning that the flow has dipped or foreign matter is in the pipes.

The future will be advanced analytics, creating accurate risk scores for all kinds of different perils across global geographies

“We are evolving to self-learning from both real and simulated data for continuous improvement and feedback into a central knowledge graph,” says Chappell. “This offers better solutions for sustainability, safety, and intelligent automation.”

Bill Peace served as deputy director of intelligence for the Serious Organised Crime Agency in the UK. Upon retiring, he became an advisor to Stop The Traffik, a coalition that aims to halt the illegal trade in human beings.

Human trafficking is a $150bn industry, with an estimated 24.9 million people in the US trapped in modern slavery, around 25 per cent of them children. However, as law enforcement struggles to follow the people, AI and data allows them to follow the money. With the help of IBM and a network of banks and authorities, Peace’s team were not only following the flow of illegal funds but also tracking communications without intruding on law-abiding citizens.

The future will see AI and machine learning used to recognise and detect specific human trafficking terms and incidents, says Peace. “AI also enables the hub to ingest open-source data — including thousands of daily news feeds — [identifying] the characteristics of human trafficking, such as recruitment and transportation.”

The road ahead

Such examples are exciting as use cases today, but this is just the beginning, and a sign of what is possible tomorrow. Thrillingly, we’re very close to achieving these smarter states. However, there are still some important obstacles to navigate on the road to progress.

While it’s not much of a leap to a smarter state, it is imperative to take the next steps carefully. Civica talks about the three “Ss” – data sharing, skills and standards. There are also ethical issues that can’t be ignored. 

There needs to be greater cooperation between industries – including the public and private sectors – to achieve the end goal. Perhaps a mindset evolution is required by many leaders to realise that collaboration rather than competition in data sharing will benefit everyone and accelerate the development of this exciting ecosystem.

This is a pivotal moment. If we collectively make the right choices, the whole of humanity can enjoy the advantages of a smarter state.

Thrillingly, we are within touching distance of enjoying data-enabled public services that empower citizens and enable better real-time decisions and smarter policy-setting by government and local authorities.

The case for opening up and connecting data has been strengthened in the last 19 months, as it is clear we all stand to gain – however, it must be done sensitively, as many of us rightly have privacy concerns.

The government and local authorities have a colossal amount of information on us all, but at the moment that data is housed in silos or duplicated. The National Data Strategy, launched in September, acknowledges that we can accelerate our public service provision by better using this data. Joining up the data will ensure the most relevant information is available when making a decision that affects the public from an individual and collective perspective.  Looking at what is within our grasp, at one end of the scale, data can be brought together to inform policy decisions and essentially drive predictive governance. Thoughts of a Minority Report-style of governing are both fanciful and impractical – and I’ll explain why in more detail below. At the other end of the scale, it is on a tactical level, delivering social care, where the most significant nationwide advantages are possible.

While the government is confident that the Strategy will catalyse the journey towards data-driven public services of the future, there are still challenges to overcome – but, encouragingly, they are far from insurmountable. Civica believes the key to unlocking the transformative power of data lies in dialling up what we call the ‘three Ss’: Standards, Skills and Sharing.

Passing ‘the flinch test’

A huge part of realising the potential of data is effective sharing. People with a high dependency on public services – for social housing, public transport, healthcare, and social care – stand to gain the most if data is shared correctly. Those in charge will have a complete picture of that individual’s needs, and the delivery of goods or services will be more frictionless, quicker, and – in theory – proactive. Putting the onus on those who might not be digitally capable is counterproductive and can lead to inefficiency and frustration. 

There is, though, a Big Brother aspect that we must consider. Taking everything the government agencies know about us and running it through an AI programme might throw up all sorts of issues. For public trust in authorities using data to be increased, the level of creepiness has to be minimised. A colleague calls this ‘the flinch test’: if you come up with an idea, and it makes you cringe, then it’s probably not going to work.

An approach to governing as explored in Minority Report – apprehending people because they might commit a crime – would likely fail the flinch test. Moreover, it would be expensive to run, as the cost of an AI false positive will be too high. Put another way, you are looking for a needle in a haystack. 

Much like deploying the use of facial recognition, we need to utilise this technology with caution. Hypothetically, if you are looking for football hooligans at a game, a false positive means that automatically resources are used to intervene – but the person only looks like someone else who is dangerous. It doesn’t take too many horror stories to drain the trust in how data is being used for public services.

Power to the people

In Scotland, the government has recently introduced citizen-centred data sharing, which means people can choose what personal information is shared. I like this idea because it hands them the power and autonomy. Still, local authorities need to explain the benefits of opening up data and why, for instance, approving data your doctor has on you being available to other services might help you down the line.

Estonia is an excellent example of how connected data can empower citizens, with 99% of governmental services available online. Some Scandinavian countries are in the chasing pack, but with a focus on data standards and skills, the UK will not be far behind.

Data sharing, skills, and standards are all interconnected, and this unholy trinity will underpin the smarter state and data-driven public service of the near future. As a confessed data geek, I’m excited by the prospect of increasing the sharing of data in the coming years. But if neither the public nor private sector develops people to have the skills to use data, it will be a massive risk in terms of missed opportunities. 

On a public service delivery level, we have to keep in mind that everyone will benefit if a council or regulator has data at their fingertips that can be leveraged to generate better and faster decisions. Getting to that point from where we are now is not a long distance, but there are complexities to overcome first – and central to progress is not failing the flinch test.

Have a range of voices around the decision table

In September, Facebook had to apologise after its AI features labelled black men in a video as “primates”. Having different faces and backgrounds in the room at every stage helps you avoid these issues, says Sudheesh Nair, chief executive of business intelligence company ThoughtSpot, who cites the UK government’s AI Labour Market research as cause for concern. “Over half of firms (53 per cent) said none of their AI employees were female, and 40 per cent said none were from ethnic minority backgrounds.” To make progress, it’s key to demystify the language around AI and data collection, he suggests. “Ensuring AI is explainable with a greater degree of transparency in training data, data gaps, and algorithmic logic can further reduce bias at scale.”

Listen to citizen journalism – scrutiny is how you improve your product

Mark Humphries, head of BI & analytics consultancy at Civica, says it’s vital that firms allow for pushback, which is playing an essential regulatory role in the early days of this tech revolution. As a rough estimate, he reckons “about 85 per cent of people shrug their shoulders” when there’s talk about a data privacy issue. Around 10 per cent are concerned, but it’s the 5 per cent who lobby for greater care and call out potential abuses that are most important. “We all owe that 5 per cent quite a debt, because they drive the scrutiny,” he says. “Having MI5 and the police have access to some of our data to keep us safe is a trade-off most people are comfortable with. But the important question is: how do you ensure that those powers aren’t abused?” Citizens who point out flaws are doing a key job in this process.

Understand when not to use AI and data

AI is only powerful when applied to the right problem in the right context. It shouldn’t be crowbarred into products where it isn’t needed, warns Dr Tim Bashford, computing portfolio director at University of Wales Trinity St David. “AI has become a buzzword – like blockchain, big data and quantum computing – presented as a panacea to every conceivable technical problem,” he says. “I recall a statistic from a few years ago that around 40 per cent of AI startup companies were, in reality, using no AI at all.” AI is a tremendously powerful tool, but it’s one of many, and certainly doesn’t fit every situation. The old joke that AI for business is like teenagers and sex – everyone talks about it, but few actually get it – is still relevant today. The message is clear: don’t use tech just for the sake of it.

AI has become a buzzword – like blockchain, big data and quantum computing – presented as a panacea to every conceivable technical problem.

Realise that not everyone welcomes the pace of change

In October, a test case in the UK courts saw one neighbour successfully sue another for £100,000 because their Ring video doorbell had breached privacy laws by filming them in their home across the street. Bashford says the greatest failures have been in rapid provisioning of ethically contentious applications without sufficient supporting legislation. “The use of AI presents ethically challenging issues with wide-reaching implications that warrant detailed discussion – which has not been taking place so far,” he says. Another recent example is the use of facial recognition technology by South Wales Police, which was found unlawful by the court of appeal. “There needs to be clear public consultancy and communication on the remit and technical capacity of these systems prior to their development and deployment,” Bashford adds.

AI and data can be a matter of life and death – respect that power

Trevor Morgan, now product manager at Comforte AG, was once involved in a well-meaning health project that applied AI to patient records to spot hidden health risks. But as it attempted to pick up on suicidal tendencies, it arguably went too far, says Morgan. “Of course, machines and software can be wrong,” he says. “It’s one thing if an impersonal machine analyses information of a sensitive nature, but it’s quite another when actual human beings obtain and use this information. When a margin of error applies to human health, you begin to question whether the technology is ripe to make fully informed decisions based on this type of machine analysis.”

The UK could build a smarter state, improving public services by connecting data from various disparate sources. This will demand greater data sharing between different branches of government – Whitehall, councils, regulators and emergency services – and collaboration between the public and private sectors. 

Working together and sharing data in an open, transparent and secure manner will drive innovation through artificial intelligence and ultimately enable and empower citizens. But progress is stalling in the private sector, due to a combination of poor data literacy at leadership level, fears of ceding a competitive advantage, and a general wariness of unintended consequences. How, then, can the public sector tap into external data sources and encourage a more collaborative spirit?

While there is no simple answer, the government’s National AI Strategy, published in late September, offers some guidance and encouragement for business leaders. The document, which sets out a 10-year plan to make the country a “global AI superpower”, is the country’s first package solely focused on AI and machine learning.

Chris Philp, digital minister at the Department for Digital, Culture, Media and Sport, is confident the publication will accelerate the development of AI and spark collaboration between public and private sectors. “We want to make sure that there are clear rules, applied ethical principles and a pro-innovation regulatory environment that can create tech powerhouses dotted across the country with the most supportive business environment in the world,” says the Croydon South MP.

We want to make sure that there are clear rules, applied ethical principles and a pro-innovation regulatory environment that can create tech powerhouses dotted across the country with the most supportive business environment in the world.

He hopes the new strategy will help narrow the skills gap to take advantage of the AI opportunity. And while data sharing is essential, standardisation is just as important; unless data is collected and managed according to common, robust rules, it might be unreliable, which directly impacts on outcomes for the citizen.

Leading by example

Matthew O’Kane is multinational IT consultant Cognizant’s global head of AI solutioning. While he welcomes the National AI Strategy, he argues that the government should take the lead in dialling up collaboration and openness.

“The government can and should set an example in the AI space by ensuring seamless data sharing across government departments,” he says. “Data is the fuel that powers AI, so through the democratisation of data across government, leaders would be able to maximise the potential to extract value from AI investments.”

Fakhar Khalid, chief scientist at Sensat, a cloud-based 3D interactive virtual engineering platform, agrees, and believes universities should open their doors, too. “A clear mindset change is needed from the top down,” he says. “Government must encourage risk in innovation and provide supportive infrastructure and resources to organisations that are willing to take such calculated risks to propel the UK as the global leader in AI innovation.”

Data is the fuel that powers AI, so through the democratisation of data across government, leaders would be able to maximise the potential to extract value from AI investments.

According to Khalid, not only are strong, open and transparent collaborative platforms needed within central and local government, but there is also an urgent requirement for more academic research to impact the public and private sectors. 

“While the government must lead the country by example, academia needs to invest more in ensuring their higher education research is fed to the industry more often than it currently does. The UK has a strong academic foundation but is slow to turn those into any commercial success.”

Chicken-and-egg scenario

Dr Mahlet Zimeta, head of public policy at the Open Data Institute, acknowledges that business leaders tend to “hoard” data, but argues that if sharing is done sensitively and sensibly, everyone stands to benefit.

“Organisations are often concerned about unanticipated use cases for their data and who might gain value from it,” she says. “They are nervous because they don’t know what business model to use. It’s difficult, as most use cases only arise when the data has been made available – it’s a chicken-and-egg scenario.”

However, Zimeta points out that there has recently been a “step change in data sharing”, with a range of industries and sectors collaborating to help the response to the coronavirus crisis. There was truly an international effort; for example, science journals changed their subscription models, allowing open access to their papers to accelerate the speed of research and development. 

“It was exciting and shows the benefits to society and the economy when more data is accessible – and as far as I know, no businesses went bankrupt as a result of making their data available.” Finally, while Zimeta calls for more cross-sector collaboration to build a smarter state, she says it’s important not to forget another potential collaborator: citizens themselves. “It’s often presented as ‘private and public’, but civil society is a crucial innovator. This data is vital, too.” She adds: “We need to start thinking about a three-way collaboration.”

Data collaborators, it’s over to you.