As digital finance accelerates, financial institutions face rising threats from increasingly sophisticated scams. In Australia, over 1.8 million people were victims of card fraud or scams in 2023, with actual numbers likely higher due to underreporting. Beyond financial losses, victims suffer emotional distress, loss of confidence, and diminished trust in institutions, underscoring the urgent need for more effective prevention strategies.
Traditional rule-based fraud detection systems remain reactive, fragmented, and prone to false positives, making them ill-suited for adaptive scams that exploit AI-generated deepfakes, synthetic identities, and social engineering tactics. This thesis aims to advance scam resilience by applying Artificial Intelligence (AI) and machine learning to identify, classify, and respond to emerging threats.
Adopting a socio-technical, human-centred perspective, the study integrates technical precision with behavioural and institutional insights. The expected outcome is an adaptive framework that enhances fraud detection accuracy, reduces financial and emotional harm, and strengthens trust in digital financial systems.