Caroline Payne, Head of Customer Advisory, SAS UK Public Sector Team
Digital innovation across governments around the world has accelerated in the last 18 months as leaders turn to data and technology to deliver rapid responses to the pandemic. Public organisations have had to move quickly, whilst being acutely aware of the need to build trust, resilience, and agility into the public services they deliver.
As a result, as we enter a period of rebuild - to build back better - specific technology trends are emerging that address critical challenges in public administration including citizen experience, ethical use of AI, digital identity, security, and fraud prevention. These areas of development all rely on access to data and have AI and predictive analytics at their heart.
At a recent UK Government virtual event which SAS hosted in September we covered several important and transformational topics that will help shape services for the future.
Understanding the connected citizen
Mike Turner explained how public sector organisations should put the citizen at the centre of digital transformation. He said, it starts with understanding why citizens share data and what they expect from the public services they engage with.
Digital channels now account for the highest volume of communications. Organisations have an opportunity to use the data they collect during digital interactions to ensure members of the public are informed, engaged and understand the services they use. Analysis of this data can also help to inform strategy and future developments.
To use this information effectively, however, organisations need to ensure they are seeing each citizen fully. A plan for communication is no longer a single broadcast event that treats everybody the same. In a connected world, the citizen themselves is key to an effective, dynamic communication processes.
By building and maintaining a rich data fabric collected from multiple sources, organisations can read each citizen’s digital ‘DNA’ (Discover, Nurture, Act) – a complex interplay of interactions which is constantly updating. It can provide an accurate picture of every individual an organisation wants to communicate with.
Once digital DNA is collected, analytical processes can add to the picture, providing more insight into who the citizen is and why they are interacting with a service. This insight then helps organisations understand what should be communicated and when. Ethical AI and machine learning techniques validate the need and recommend the proposed response.
Ethically distilling insights from the nation’s digital DNA will drive accuracy and effectiveness of interactions – and through better engagement with citizens, help to create a fairer, safer and stronger public sector.
Protecting connected citizens from fraud and identity theft
As the volume of digital interactions increases and public sector organisations develop better mechanisms for digital engagement, the obligation to protect the identity of those citizens becomes paramount. Three SAS colleagues spoke passionately about their experiences with customers – here Colin Gray, Glenn Smith and Weder Souza tell us more.
Scams and social engineering are a rising threat with fraudsters engaging ever more sophisticated methods. Global crises have a strong correlation with fraud activity and all organisations need to be vigilant in the wake of the pandemic. Estimates of fraud in the UK public sector vary widely, from between £30bn-£40bn to over £100bn[1]. Identity theft is growing 10-15 percent year on year[2]. Covid hasn’t done anything to diminish these numbers.
What can public sector organisations learn from the financial services sector where identity checks and the ability to ‘know your client’ are critical to understanding who you are doing business with?
In Financial Services, as in the public sector, there has been a rapid shift away from in person interactions towards digital engagement which can make it harder for to verify and authenticate people’s identity. While there is a fine balance between detecting and preventing fraud, it is possible to make processes easier for more vulnerable customers whilst impeding the progress of higher risk cases.
If we can start by establishing a person’s identity, and, apply the appropriate level of friction in the engagement process, we can better avoid identity theft. Organisations need to use data and analytics to effectively summarise the risks and then act swiftly, allowing or denying the user to perform a task or selecting an event to be further investigated. The main objective is to take a better decision based on the integration of all data than would have been possible with data in different silos.
UK Finance is now calling on the UK Government to include economic crime in the forthcoming regulatory framework on online harms for greater protection to individuals. Incorporating economic crime within this regulatory framework will help to harness the capabilities, expertise, and powers of both the public and private sectors, to truly create a step change in approach.
Building trust through Responsible AI
As AI expands its reach into more applications, public interest in the topic is increasing and awareness of ethical issues in the use of AI is rising. While some stories involving AI may be
oversimplified in the press, public sector organisations need to put effort into using AI in the right way or risk being associated with unfair or discriminatory practices. Iain Brown is our expert.
All public sector organisations using AI need to be asking themselves the following questions:
- Do you know what your AI is doing?
- Can you explain it to members of the public?
- Will they respond happily when you explain it?
Human beings are subject to multiple sources of bias and algorithms are only as good as the data that fires them. Eliminating bias and ensuring transparency, privacy, reliability, and inclusiveness in AI are essential for fostering trust with the public.
GDPR and an expanding number of initiatives and frameworks are helping to establish ethical AI standards as a minimum requirement for all organisations. A recent survey by Accenture showed that 63 per cent of CEOs have established oversight committees and further 13 per cent are considering it, to ensure accountability and avoid bias creeping into AI applications.
Implementing a strong governance framework is an essential step in ensuring an organisation’s data is being used to create beneficial outcomes.
Putting AI into practice
As more organisations recognise the important role of AI in digital transformation, what are the major considerations to ensure successful application of the latest technological advances?
Balancing algorithms with outcomes
Janice Newell had some fundamental tips. Always consider usability versus complexity. It can be tempting to deploy the latest developments in data science and make use of opensource techniques that are being constantly updated. However, organisations need to make sure they are not limiting their thinking when it comes to algorithms and always have the outcomes in mind. Think about the time taken to develop and deploy models and whether that is appropriate for the problem you are attempting to solve.
Likewise, an algorithm on its own doesn’t solve a business problem. Business logic needs to be applied alongside the algorithm – for example what score will be used to define high, medium and low risk? Having the ability to test, build and integrate rules with models reduces error and time to deploy. Time to value is an essential consideration.
Considering the bigger picture
It is always worth taking a step back and considering where AI can best help in the whole digital transformation journey. It’s important to think about end-to-end processes, how assets can be reused and how will you provide a consistent, citizen experience. Better planning and design up front may take more time initially, but will improve efficiency, prevent rework and reduce costs in the long run.
Strengthen the supporting framework
What can you do to ensure you have the right processes in place and are using technology in the right way to reduce costs over time? The typical tenure of a data scientist is 18 months, so organisations need to set up processes and standards that enable them to capture that knowledge and maintain transparency for continuity. An organisation’s people, processes and technology need to work in harmony to support the analytics lifecycle, incorporating automation where possible.
As you can tell there were some interesting topics discussed all which will accelerate the UK forward as the UK government continues it digital transformation journey. In addition, we were delighted to have a conversation with DWP, HMRC and our colleagues from Microsoft, you can read more in our blog Resetting the Risk Dial.
To learn more about AI & Analytics for Government Innovation at SAS visit our webpage.
