by Ashwini Singh
Being one of the buzz phrases of the year, Artificial Intelligence has assimilated itself into numerous sectors of academia and business. In the forensic investigation and auditing spheres, Artificial Intelligence has proven that it can be used as an invaluable type of Open Source Intelligence to audit a subject’s digital footprint.
Digital footprints, lifestyle audits, and Open Source Intelligence
Kaspersky describes a digital footprint as the trail of data a person leaves behind whenever they use the internet. Essentially, any information submitted online (such as emails, messages, and the like) contributes to a person’s digital footprint. Considering that the United Nations recently reported more than eight billion people in the world (five billion of which have internet access as of 2024), there are literally billions of digital footprints on the World Wide Web that can serve as a treasure trove of information when conducting a lifestyle audit.
Law and commerce academics Jacqui-Lyn McIntyre, Duane Aslette, and Nico Buitendag have best defined a lifestyle audit as “a process of tests performed on a person’s financial information to determine if the person’s lifestyle corresponds with his/her known sources of income”. Ultimately, the goal of a lifestyle audit is to determine whether a person is living above their means, potentially hinting at the use of illicit funds if the person’s lifestyle is beyond their known income. As a result, a person’s digital footprint can leave behind crucial clues about their lifestyle which – in comparison with their known assets – can be used to determine if the person has undeclared funds that require further investigation.
According to the SANS Institute, Open Source Intelligence (commonly referred to as “OSINT”) is intelligence gathered via publicly available information and analysed for a specific purpose. Traditional OSINT tools include online public records, social networks, search engines, etcetera. The main commonality among traditional OSINT tools being that information had to be manually searched for by the investigator. However, with Artificial Intelligence (AI), the searching and retrieving of information can become a completely automated process.
Artificial Intelligence as an Open Source Intelligence tool
AI, being primarily computing that emulates human intelligence, experienced a boom in its use in recent years. Microsoft reported that ChatGPT (a popular generative AI chatbot) reached 100 million users in just the course of two months, with AI usage by enterprises increasing to 75% in 2024 from 55% usage the previous year. Evidently, the adoption of AI is exponentially growing.
A 2023 study published by Springer discussed the use of AI-powered OSINT tools, noting that AI automates the process of collection and analysis of digital data. In the context of a forensic audit, a practical use for AI as an OSINT tool arises in the collection of data about an investigation subject.
AI tools such as Google’s Gemini and Microsoft’s Copilot contain deep research features that are capable of extensively digging through search engines to obtain information about a search query. In the instance of a search query requesting information about an investigation subject as part of a lifestyle audit, these AI tools can perform the functions of OSINT tools by soliciting the data about a subject’s digital footprint, namely business records and social media posts that demonstrate the lifestyle of the subject (which can be compared with their known sources of income).
For example, take the circumstance of an investigator auditing a subject suspected of concealing foreign income to evade taxes. The investigator can request that an AI tool (like Copilot or Gemini) retrieves information about an investigation subject’s business records around the world. The AI tool, accessing OSINT tools, can retrieve public records that indicate information about the various companies the investigation subject is a principal of – a function which is usually performed in a span of minutes with modern computing technology. With the data retrieved via an AI tool, the investigator can follow these leads to ascertain the details of any undisclosed income accrued via the businesses belonging to the investigation subject.
The applicability of AI tools further extends to eliciting details about an investigation subject’s social media accounts, where (subject to privacy settings) a subject’s posts can be viewed to understand if there are any noticeable red flags in relation to their expenditure activities posted online (left behind by their digital footprint). To demonstrate, an entry-level desk admin officer earning minimum wage posting photos of an expensive vehicle and claiming to have bought the vehicle by cash would definitively raise suspicions as to where the funds for the vehicle came from. AI tools can bring to the attention of an investigator, who perhaps is conducting a lifestyle audit on the admin officer, that the admin officer made posts online that the investigator would need to look into.
While AI has the potential to become an indispensable productivity tool for investigators conducting a lifestyle audit, AI tools in their present state have shortcomings that need to be considered before engaging them.
Moving forward with Artificial Intelligence in investigation
As the world embraces the 5th Industrial Revolution, marked by automation and AI, the Massachusetts Institute of Technology has cautioned that AI tools maintain issues with their accuracy. Consequently, the onus rests on any person who uses an AI tool to verify the information generated by the AI for its factual accuracy. As would be the case with using OSINT tools, any investigator must vet data about an investigation subject’s digital footprint (obtained via an AI tool) with care and scrutinize the information for corroborative evidence that can prove its accuracy. Conclusively, AI – although possessing a degree of intelligence – cannot replace the human qualities of scepticism and logical reasoning that make a sound investigator. While Artificial Intelligence is an interesting assistive tool for auditing digital footprints, it must be used with a grain of salt.
Short Author Bio:
Ashwini Singh is a multidisciplinary author and academic whose work is featured in The Star, The Witness, and the Cape Argus. A 15-year expat in the UAE and alumna of the American Community School of Abu Dhabi, she earned qualifications in Medical Law (LL.M.), Forensic Investigation & Criminal Justice (PG Dip.), Communication Science (B.A.), and TESOL. She is also an affiliate of ACFE® South Africa and an associate of the Professional Editors’ Guild.
