Government productivity and transparency are hot topics, with trust in public institutions declining worldwide and global public debt levels nearing 100 percent of global gross domestic product.

Managing fraud, waste and abuse (FWA) is key to public sector productivity. The British Government estimates that £39.8 billion to £58.5 billion of public money is subject to fraud and error yearly. In the US, cumulative improper payment estimates by executive branch agencies have totaled about $2.8 trillion since fiscal year 2003.

Monies lost to fraud and improper payments are often difficult – and sometimes impossible – to recover. Public programs may suffer as a result of this lost funding. Yet, the taxpayer ultimately loses. More money is then allocated to cover the shortfall, budgets are increased, and taxes are raised, or deficit spending ensues.

Tackling fraud in the public sector

Many government agencies are looking to technology, specifically data and AI, to help with budget deficits, reduce FWA, and regain public trust. Where the government's focus was once on FWA detection and punishment of offenders, in some cases, it has shifted to helping identify errors and compliance issues before penalties are administered.

For instance, Malta Tax and Customs Administration (MTCA) recognized that the public saw them as punitive, relentlessly pursuing tax offenders. The MTCA is changing the minds of the Maltese through organizational transformation and proactive engagement with its citizens. The administration knew building trust and changing public opinion starts with improving customer service.

“We are not investigative officers in pursuit of culprits, but an administration that can assist citizens and businesses in adhering to regulations, and we can be a valuable partner in supporting taxpayers,” said Joseph Caruana, Commissioner for Tax and Customs at Malta Tax and Customs Administration.

Three key actions that MTCA took that was effective and improved public favorability:

1. Enhanced transparency in tax policy.
2. Strengthened legal frameworks to ensure fairness and compliance.
3. Adopted new technologies for better data management and analysis.

Break down barriers with data and AI

Whether focusing on FWA prevention or detection, getting accurate, complete and timely data is essential. My SAS colleagues Carl Hammersburg, Ensley Tan and Stephan Goddé discussed in a recent webinar, Innovate to Eliminate: Using Tech to Fight Fraud in the Public Sector, that technological and cultural barriers can prevent public sector bodies from getting the data they need when they need it.

They outlined the following challenges:

• Legacy data silos and technologies: Some systems aren’t integrated and don’t easily talk to each other – if at all. Understanding what data is available to support the FWA reduction effort, let alone getting the data, may be too time-consuming, resource-intensive, or deemed not worth the effort.
• Volumes of raw and unstructured data: Some localities still heavily rely on hard copy documents and unstructured formats like PDF images. Documents may need to be converted to OCR-readable formats and, in some cases, translated to the local language or English before use in FWA detection and prevention efforts.
• Data restrictions: Laws, regulations or norms often govern how different datasets are accessed or used. In protecting the organization, legal, privacy, and other internal teams may unintentionally hinder FWA detection and prevention efforts without clearly understanding the exact restrictions for a particular dataset.

So, what can the public sector do to address these challenges? My colleagues unanimously agree that the number one thing to do is to gather the owners of the underlying business issues together. These individuals gain the most from the successful use of data and AI, and they can identify the must-have data, who needs it and how they need to use it.

The promise of AI for payment integrity

While generative AI (GenAI) in the public sector is relatively new, AI has proven extremely useful in tackling situations with large data volumes, short timelines and significant complexity. These programs and services are seeing benefits:

Public sector procurement: A single government agency may have hundreds of thousands of suppliers, millions of items received, billions or trillions of invoices and payments, and even more pages of contracts. Without AI, detecting supplier collusion and duplicate invoices or negotiating better pricing would be difficult.

Tax compliance: Identifying and prioritizing potential compliance issues across businesses' tax returns is a big undertaking. With AI, the time spent in triage and review could be reduced, allowing more focus on making data-informed decisions that drive value for your agency.

Social benefits: Nutrition assistance program fraud hurts our most vulnerable. There are numerous types of fraud – from skimming to false applications to multiple applications, for instance. With AI, scanned document images become structured data, usable formats and more accessible to a wide range of users. Digital transformation is a win for the agency and citizens who need services.

According to my webinar colleagues, GenAI in the public sector has been limited to “less risky” projects such as internal search engines or internal ideation tools. The use of GenAI for FWA detection and investigations still seems to be some time away.

Use data and AI to fight AI-enabled fraud

The fight against FWA in the public sector is evolving rapidly. A recent global survey by a third-party research firm reveals data and AI's potential for fighting fraudulent activities and optimism about the future of fraud detection and prevention.

The public sector must embrace data and AI, innovating FWA detection, investigation and prevention methods. Moving from legacy systems to data and AI can improve public sector productivity, transparency and accountability, paving the way for restoring public trust.