Unveiling the Synergy of OSINT and AI in Fraud Investigation 

In the rapidly evolving landscape of fraud investigation, the convergence of traditional Open-Source Intelligence (OSINT) techniques and groundbreaking Artificial Intelligence (AI) tools is paving the way for a new era of digital sleuthing. At the 2024 ACFE Fraud Conference Europe, in the session "Fraud Investigation in a Digital Age: Using Traditional OSINT and AI Tools", speaker Stephen Hill, PhD., CFE, provided attendees with invaluable insights into how this blend of old and new methodologies is revolutionizing the field.  

The Backbone of Modern Fraud Investigation 

At its core, the session underscored the indispensability of traditional OSINT techniques in fraud investigations. OSINT, with its roots in gathering intelligence from publicly available sources, has long been a cornerstone of investigative work. The digital age has expanded the scope of OSINT to include a wealth of online data, ranging from social media footprints to data breaches. This session highlighted how these traditional methods remain highly effective in unmasking fraudulent activities by providing a solid foundation for any investigation. Hill spoke about the use of social media and digital media in modern fraud cases through extracting data about fraudsters actions and whereabouts. He also explained how some AI social media tools can be used for automating data collection, analyzing large volumes of data and identifying patterns or anomalies. 

Using ChatGPT in Fraud Investigations 

Hill spoke about the usage of ChatGPT in fraud investigations, stressing the normal concerns over the accuracy of data and touching on the importance of not inputting proprietary or confidential information. He spoke about using ChatGPT as a search engine, asking questions on how to find certain information on the internet or where to look to access certain records that may be relevant in an investigation. Hill also spoke about Google Gemini (formerly Bard) and how it can make up for ChatGPT’s shortcomings, specifically that it can leverage up-to-date data where ChatGPT is limited to information pre-2021.  

A Dual-Approach Strategy 

One of Hill’s key takeaways was the importance of integrating traditional OSINT techniques with AI tools to form a robust fraud investigation strategy through suggestions like the ones above, as well as using AI to help detect potentially AI-created content or deepfakes. This dual approach leverages the strengths of both methodologies—OSINT's depth and breadth of data, combined with AI's speed and analytical capabilities. Attendees learned how this integrated strategy not only saves time and resources but can also significantly improve the outcomes of investigations.