by Jasmine Kendall and Sulamaan Rahim
At Oxford Insights and Butterfly Data, we are interested in finding new ways to tackle corruption and foster open, transparent governance. Oxford Insights have conducted extensive work with the Open Contracting Partnership, Open Innovation, and the Extractive Industries Transparency Initiative (EITI) on research into global procurement responses during COVID-19 and evaluating programmes promoting transparency in the extractives industry. Butterfly Data has extensive experience in compliance risking, specifically in identifying anomalous claims and helping to claw back overpayments from COVID support schemes such as Eat Out to Help Out. Together, we’re partnering on a machine learning tool which will allow users to identify corruption risk in public contracts.
Tackling COVID corruption and recouping tax money has also been at the forefront of the new Labour government’s agenda, with Rachel Reeves making a number of announcements in advance of the budget. In July, she announced she would be appointing a COVID corruption tsar. She has also more recently announced investigations into £600m worth of COVID contracts.
Corruption is a persistent problem globally that manifests in many different ways. Though high-level global estimates are difficult to substantiate, it is plausible that there are costs in the trillions of dollars globally. Curbing it is therefore vital: it has fiscal ramifications beyond just the sum of lost tax revenue, for example. IMF research suggests that corruption not only diminishes tax revenue but distorts public spending in such a way that public money drains away from social goods like education, healthcare, and infrastructure.
Such behaviour became more prevalent during COVID as emergency response protocols opened the door for procurement deals that did not have to undergo stringent checks. A House of Commons report from 2023 found that ‘since the beginning of the COVID-19 pandemic, the level of fraud against the taxpayer has increased fourfold’ in the UK. Moreover, Transparency International found corruption red flags in £15.3bn worth of COVID contracts following analysis of public procurement data.
What can AI do?
When it comes to exactly how to recoup these losses, many have suggested the application of artificial intelligence to both proactively prevent and recoup these losses. Governments around the world are piloting projects using AI and machine learning models to more effectively identify procurement ‘red flags’.
For example, in Brazil, Alice, a tender analysis platform uses AI to analyse data from various procurement portals in order to deliver real time risk alerts to public servants. Responding to a recent survey of 26 EU countries, seven countries responded that they are already using AI to analyse their procurement data, or help prepare tender documentation, whilst another 10 countries have plans to do so.
However, for these AI solutions to be truly accurate and impactful, high-quality data is essential. Without a strong foundation of reliable data, AI-driven insights can easily become flawed or biassed, weakening efforts to detect and prevent corruption. Encouragingly, AI also holds promise in improving data quality itself: authorities in Romania, Bulgaria, and Portugal are planning to pilot AI-based initiatives aimed at enhancing their data integrity.
Though still nascent, AI technologies have the potential to curb corruption’s prevalence and mitigate some of the problems that it causes. The vast amounts of data that AI technologies can analyse and identify patterns amongst make public procurement data, for example, a very promising use case. Corruption, however, cannot have a solely technical solution. There must be political action and will: data must be used to serve legal and civic action and broader movements for open and transparent procurement processes must be undertaken alongside, and perhaps using, these tools.
Looking forward: from theory to practice
At Oxford Insights and Butterfly Data we are keen to contribute to the growing body of case studies in this space and deepen our collaboration through the development of a machine learning tool designed to identify corruption risk in the UK’s COVID-19 procurement. With this tool, we aim to move beyond theory, and pilot a practical and actionable solution which directly addresses losses due to corruption.
We will be providing updates as the initiative develops, but please get in touch with us at info@oxfordinsights if you want to discuss further!
Photo by Matt Brown on Flickr
