AI-Powered Investigations: The Future of Exposing Crime

Today, I realised how dramatically AI can help in investigating criminal activities through the analysis of publicly available information. When I think about it, investigators, journalists, and state authorities have always relied on their ability to connect information, recognise patterns, and uncover hidden links. But until recently, this was purely human work—slow, exhausting, and dependent on a person’s ability to notice details and combine facts. AI can speed up this process and take it to an entirely new level.

Crime Investigation in the Age of Artificial Intelligence

Take business registries, for example. Many countries now have them publicly accessible, and we can search them manually. But what if AI could automatically scan these registries, compare ownership structures, and detect suspicious patterns? What if it could identify that a certain company has an unusually high volume of transactions with another, which is linked to a politically exposed person? Or track how company ownership changes shortly before securing a major government contract? This is already possible today, yet it is not being fully utilised.

The same applies to public procurement. If the same companies repeatedly bid for tenders, if suspiciously similar pricing offers appear, or if an unknown company with minimal history unexpectedly wins, AI could quickly flag these irregularities. People might notice such patterns by chance, but a system that is continuously running and analysing all available data could detect them instantly.

Another fascinating use case is AI's ability to monitor the wealth of politicians and business figures. Today, asset declarations exist, but few systematically compare them with reality. What if AI could automatically pull data from land registries, corporate financial statements, and asset disclosures to identify discrepancies? For example, when someone with an officially low income owns luxury real estate or when expensive properties are purchased through a network of obscure companies.

Then there are social networks. Public posts, photos, location tags—all these are small puzzle pieces that can be assembled into a picture that someone never intended to reveal. AI could identify that two individuals who claim not to know each other actually spent a vacation together. Or that a politician who insists they have never met a particular business owner was in fact present at their private celebration. It’s both terrifying and fascinating how much can be reconstructed from publicly available data.

It’s not just about politics and corruption. AI can be extremely useful in investigating tax fraud as well. Large-scale money laundering schemes often involve funds moving through dozens of companies in different countries. A human analyst might spend months mapping out these transactions, but AI could accomplish this within minutes. It could track money flows, analyse blockchain transactions, and detect patterns indicative of money laundering.

The Risks of AI Being Used for Extortion and Control

I also reflect on what impact this could have on society. On one hand, there is enormous potential for uncovering fraud and criminal activity. On the other, there is the issue of privacy and the misuse of power.