An AI LLM Just Ran Its Own Cyber-Espionage — and Africa Was Locked Out of the Next Frontier

Mikael Alemu Gorsky's paper, submitted to arXiv in June 2026, examines two landmark events that define AI's role in cybersecurity. In 2025, a large language model autonomously executed the vast majority of a state-aligned cyber-espionage campaign, with human operators intervening only at a few decision points. In 2026, the most capable cyber-relevant model was placed under a controlled-access program limited to a vetted set of US technology firms, allied governments, and European standards bodies—explicitly excluding any African government, operator, or university. Together, these events prove that frontier language models are now decisive instruments of cyber operations, built, owned, and rationed within a small circle from which Africa is absent.

The paper quantifies Africa's exclusion across three axes: skilled people, compute and electrical power, and investment—all measured against current figures. Meanwhile, AI-enabled fraud is already mounting against African mobile-money systems, a sector where the continent leads globally. Because comparable but ungated models are forecast to spread within six to twelve months, Gorsky argues for a rapid, African-led response. The recommended approach includes threat-intelligence sharing, adoption of governance frameworks, and strategic partnerships—all operating inside that narrow window to mitigate the risk of a permanently dependent digital economy.