Newest papers

Make the Adversary Spend More Energy: Generative Deception against Language-Model Attackers in Critical Infrastructure

Vlad Mihal, Hosam Alamleh, Alessandro Cantelli‐Forti · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.27739 · cited by 0 works (OpenAlex) · open-access PDF

Portuguese

Os atacantes do modelo de linguagem consomem recursos computacionais à medida que exploram e interagem com um alvo, criando uma oportunidade para a fraude cibernética impor custos. Se tais compromissos favorecem o defensor requerem a medição de ambos os lados. Nós avaliamos um atacante do modelo de linguagem autônomo contra um decoy que usa um modelo de linguagem defensiva para analisar comandos do atacante e selecionar respostas defensivas. Ao longo de 330 sessões, oito pares de modelos, dois perfis de atacante e três níveis de contra-injeção, ambos os lados são executados em GPUs idênticos e sua energia é medida acima do vazio. Com o pensamento em tempo de execução desativado, o token de defesa-atacante tem um intervalo de energia 95% acima de um em

Machine translation (Meta M2M100); the English original follows.

Abstract

Language-model attackers consume computational resources as they explore and interact with a target, creating an opportunity for cyber deception to impose costs. Whether such engagements favor the defender requires measuring both sides. We evaluate an autonomous language-model attacker against a decoy that uses a defending language model to analyze attacker commands and select defensive responses. Across 330 sessions, eight model pairs, two attacker profiles and three counter-injection levels, both sides run on identical GPUs and their energy is measured above idle. With runtime thinking disabled, the attacker-to-defender energy ratio has a 95% interval above one in 41 of 48 configurations under an allocation excluding attacker diagnostic calls, and in all 48 when those calls are included. Token ratios remain within approximately a factor of two in configurations without generation-cap truncation, showing that token counts can obscure substantial energy asymmetry. Counter-injection increases attacker tokens by approximately 9% per level without a detectable pooled change in energy consumption. Enabling thinking on both sides increases defender energy by 8--17 times and reverses the exchange for one model pair in the original campaign. We use these findings to motivate a proposed SCADA deception architecture with private physical state per identified adversary and deterministic protocol and process emulation.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Adversary, Computer security, Security token, Computer science, Deception, Process (computing)

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange