The main challenges the UK government faces in using data analytics to tackle fraud have been outlined, along with a consultation launched to further explore how these issues can be addressed.
The Cabinet Office’s Counter Fraud Centre of Expertise and the Department for Digital, Culture, Media and Sport (DCMS) has launched a thought paper, Tackling fraud in government with data analytics: Starting the conversation, which has identified where and how government departments can use data to prevent and identify fraud.
Private sector and academia have contributed to the document, which is the result of four years of work aimed at tackling fraud and error loss, estimated to cost the government £31-£49bn yearly.
“Opening up data in a way that makes it reusable and easily accessible, while taking into account legal and ethical considerations, can deliver a number of positive benefits for the government, citizens and the economy,” said digital minister Margot James.
Intended to “help build better citizen understanding of government data sharing, and bring to life the value and importance of data”, the paper provided insight into joint work led by the Cabinet Office and the DCMS to advance use of data to counter fraud across public sector, outlined key hurdles and called for input.
As part of intentions to make better use of data to drive savings and improve citizen service delivery, the government is also looking to create a National Data Strategy and an open call for evidence has been launched in June to support the process.
According to the paper on the role of data in tackling fraud, the government has recently changed the way it views fraud, and there is an acceptance that economic crime is exceptionally difficult to measure. Another realisation is that more proactivity is needed to look for and find fraud in individual departments.
The Centre of Expertise for Counter Fraud is working on building an evidence base and a single picture of the nature and scale of fraud across departments. According to the paper, this is expected to help central and local government bodies tackle the issue.
However, many challenges that could hamper progress in the use of data analytics to tackle fraud have been outlined in the paper (see box).
A key hurdle is related to data access between parties. According to the paper, the government sees benefits in working with private sector in areas such as car insurance fraud, while technologies such as distributed ledger or blockchain could make it easier for departments to share data without the need for continually re-requesting it.
Work done to address the issue of data sharing so far include the ongoing development of a cross-government Counter Fraud Function to enable sharing and analysis of data, with the Centre of Expertise “investing in providing tools and legislation” to make it work.
This investment, the paper notes, includes the introduction of rules to provide better access to data, and data services and exploring sharing frameworks, so that public bodies can share data with one another in a simplified manner while ensuring legal compliance.
The introduction of an approach to quick testing of analytical techniques, with reusable components for tackling fraud problems across government, as well as the development and delivery of best practice guidance, is also expected to be introduced as part of the initiatives under the wider function.
A governance structure, the Counter Fraud Data Alliance, will be managed by the function to provide a framework for data sharing across public bodies and the private sector “to test the value of such collaborations”.
Ethics is another key challenge raised in tackling fraud. According to the paper, the government realises the potential of artificial intelligence (AI), but also understands the implications of its use are complex and far-reaching.
The Centre for Data Ethics and Innovation has started its activities earlier this year to identify the measures needed to strengthen and improve the way data and AI is applied in government in an ethical manner, drawing on insights from regulators, academia, the public and businesses.
However, when it comes to tackling fraud, the government wants to understand to what extent should the government seek to pursue use of technologies like AI to tackle fraud, and what frameworks should be put in place to ensure their usage is ethical.
Improving data quality is another issue raised in the paper, and this particular area encompasses technical and more ethical challenges.
This includes means and processes through which the sources of data will be vetted, as well as decisions that can or should be made to constitute the official version of the data, as well as moral and ethical limitations of how that data will be used by the government, business or citizens.
Ensuring sufficient data capabilities is another challenge outlined in the paper, so ensuring public sector staff have relevant skills and capabilities to generate maximum value from using data to reduce fraud.
Ongoing initiatives to tackle that particular challenge include the coordination of a new Counter Fraud Data Analyst Community, aimed at sharing learnings around countering economic crime across government with data analytics.
Developing a “data mindset”
Another hurdle listed is to do with developing a “data mindset”, getting individuals to understand the value of information, which then impacts how policies and processes are designed.
To get answers around questions surrounding the challenges outlined in the paper, the Cabinet Office is looking for input from citizens, government, industry and academia, which can be submitted through a form or an online survey available through the paper.
The five challenges in using data to tackle fraud in government
- How – and if – a data mindset should be embedded in departments;
- How can data quality be improved, with data sets consistently used and understood;
- How can staff be skilled to use data efficiently to tackle fraud;
- How can access to data be improved to counter fraud effectively;
- How data-driven innovations can be used to tackle fraud and what frameworks need to be introduced to ensure ethical use.