AzureFeeds

Chief Technology Officer and Microsoft Regional Director, Dr David Goad, recently highlighted the transformative potential of generative AI in combating financial scams, in a recent live stream on the Microsoft Reactor YouTube channel.  

With a wealth of experience in artificial intelligence and machine learning, David shared his expertise and insights, providing practical examples on the applications of generative AI and Azure AI Foundry for scam and fraud detection in the banking sector. 

The rise of digital banking and fraud 

David Goad began by setting the scene, discussing the significant shift towards digital banking over the past decade. This transition has brought numerous benefits, including convenience and access to information, but it has also led to a surge in financial scams and fraud. The banking industry, particularly in Australia, has seen billions of dollars lost annually due to these fraudulent activities.

Key industry trends 

David highlighted several key trends and challenges in managing fraud and scams within the banking sector. He pointed out that phishing remains one of the most prevalent methods used by fraudsters, with targeted spear phishing and whale phishing aimed at senior individuals. This demographic is often targeted, leading to significant financial losses and stress.  

Additionally, phone calls, text messages, and emails are common delivery methods for scams.

Opportunities for improvement with generative AI 

David Goad emphasized the opportunities for improvement in the scam detection process through the use of generative AI. He explained that generative AI can enhance various aspects of fraud detection, including identifying fraudulent emails and texts, summarizing customer complaints, categorizing complaints for efficient routing, and evaluating customer sentiment. By leveraging generative AI, banks can improve the accuracy and efficiency of their fraud detection processes, ultimately reducing customer frustration and enhancing service levels.

Demonstrating generative AI for phishing detection 

To illustrate the practical application of generative AI, David demonstrated Azure AI Foundry and Azure OpenAI Studio. He showcased how generative AI can be trained to identify phishing emails by fine-tuning a model with a dataset of classified emails. The model that David presented, was able to classify emails as phishing or non-phishing and provide explanations for its decisions, demonstrating the potential for generative AI to streamline the fraud detection process. 

Learn more 

David Goad’s presentation emphasized the transformative potential of generative AI in the banking sector and the opportunities that adopting generative AI can bring to help customer experiences. For those interested in learning more about this topic, David has written a LinkedIn article that delves deeper into the use of generative AI in fraud detection.

To watch the full recording of David Goad’s insightful presentation and technical demonstrations, visit the Microsoft Reactor YouTube channel.