Investigative AI: Legal and Illegal Economies. Investigative AI: Intelligence Topology of Terroristic Attacks
May 20, 2026 — 11:45 am - 12:45 pm
Maxi Stage 1
Already registered? Log in now to personalize your experience!
Description
This presentation introduces an Investigative AI framework built on artificial neural networks
developed at the Semeion Research Center (Rome), specifically AutoCM (Auto-Contractive Map) and Meta-AutoCM. The analysis is applied to a dataset of 120 monthly observations (2016–2025) across 32 heterogeneous variables spanning corporate profits, energy markets, military expenditure, illegal markets, criminal trafficking, geopolitics, financial opacity, and cognitive-political dynamics (social media, polarization, disinformation). The work distinguishes two prediction paradigms — probabilistic (classical pattern recognition) and harmonic (weak present-day signals that foreshadow major future events) — and presents a three-stage analytical pipeline: PCA for initial orientation, AutoCM for first-order topology ("who is linked to whom";), and Meta-AutoCM for second-order functional topology ("who acts like whom"). Key findings show that the variables with the highest Activation Force in the system are not the large legal economic actors but rather illegal or extreme nodes — fentanyl (Act = 1.000), cocaine, terrorism, diamonds — which operate as high-perturbation triggers with few connections but strategic positioning. Legal and institutional nodes (Meta Platforms, cryptocurrency, Lockheed, social media) dominate connectivity and serve as signal amplifiers and redistributors. AutoCM reveals a concrete structural corridor: drugs → diamonds → energy → refugees → military apparatus. Meta-AutoCM uncovers unexpected
functional equivalences (Meta ≈ Crypto ≈ Lockheed) and demonstrates that JPMorgan is peripheral, with zero activation. A bootstrap validation across five 75%-subsamples confirms the granite stability of the trigger core and the fluidity of the redistribution apparatus. The central thesis is that economic volume does not equate to structural power: systemic control arises from the
coupling of a few illegal ignition points with legal amplification corridors.
Semeion directs its research towards the conceptualization and design of Artificial Organisms composed of adaptive systems based on Artificial Neural Networks and Evolutionary Algorithms for the...
