Evil-AI Benchmark is introduced, an open-source evaluation framework for assessing LLM agents in smart environments through five adversarial threat vectors—prompt injection, persuasion, attacker-in-the-middle (AITM) simulation, data leakage, and unsafe actions—and a capability validation that checks whether a model can correctly activate device-control tools.
Abstract
Smart environments increasingly integrate artificial intelligence (AI) to process sensor data, coordinate Internet of Things (IoT) devices, and actuate cyber-physical services. Within these environments, large language models (LLMs) are emerging as decision-making and interaction components. However, LLM integration expands the attack surface of these environments, and existing benchmarks focus largely on risks associated with text-based interactions and LLM-generated outputs, failing to capture cyber-physical risks that arise when models control physical actuators. This paper introduces Evil-AI Benchmark, an open-source evaluation framework for assessing LLM agents in smart environments through (i) five adversarial threat vectors—prompt injection, persuasion, attacker-in-the-middle (AITM) simulation, data leakage, and unsafe actions—and (ii) a capability validation that checks whether a model can correctly activate device-control tools. The benchmark instruments tool activations and forwards tool-trigger events to an external Arduino-driven servo actuator, providing physical verification of actuation events. To demonstrate the benchmark in practice, we evaluate eight representative LLMs using 350 executable scenarios: 250 adversarial tests spanning five smart-environment domains (smart home, healthcare IoT, industrial control, public infrastructure, and smart building), 50 capability checks, and 50 benign over-refusal tests. Responses are scored by two independent judge models and verified by a human evaluator. We report Evilness as the extent to which attacks succeed. The Evilness Score equals the number of successful attacks, and the Evilness Rate is the corresponding percentage. Across models, Evilness Rates range from 0.8% to 54.0% (2–135 successful attacks out of 250), while Defense Rates range from 46.0% to 99.2%. The most persistent weaknesses occur in persuasion, especially under sustained multi-turn pressure, and in prompt injection. Moreover, larger models are not necessarily more robust against adversarial attacks. We additionally quantify over-refusal on benign authorized requests, so that safety is not rewarded by indiscriminate refusal. Evil-AI Benchmark provides reproducible diagnostics and quantitative metrics that connect text-based safety evaluation with real-world cyber-physical operation.
AI-enabled Cyber-Physical Systems (CPS) are highly vulnerable to adversarial and anomalous inputs, where small perturbations can induce cascading errors and unsafe control actions. Existing approaches, such as rule-based filtering, training-time regularization, or diffusion-based reconstruction, either operate outside...
Zi-Yan An, J. Stankovic, Mei-Yi Ma· arXiv.org· 0 citations
Artificial intelligence agents increasingly automate cybersecurity tasks, including offensive operations. Frequent attacks across the web make endpoint detection and response (EDR) a critical defense, but whether AI agents can evade it remains unclear. Benchmarks evaluate the offensive capabilities of large language mo...
Wei Yuan, Zi-Fan Chen, Yu-Cheng Zhang et al.· 0 citations
The paper is an agentic artificial intelligence architecture of Security Operations Center (SOC) automation that meets the challenges of signature based intrusion detection to large scale, multi-vector cyber threats. Current SOC systems have deficiencies in alert fatigue, absence of contextual reasoning, and uncontroll...
Varun Bindra, Ankit Dubey· 2026 International Conferenc...· 0 citations
The emergence of agentic artificial intelligence (AI)—systems capable of autonomous decision-making, goal-directed action, and interaction with complex digital and physical environments—represents a fundamental shift in AI. Securing these systems requires moving beyond the protection of individual models to the protect...
Elisa Bertino· Proceedings of the 2026 ACM...· 0 citations
Security Operations Centers (SOCs) for information technology and operational technology share one incident-response problem: a flood of correlated alerts and too few analysts. Large Language Models (LLMs) are increasingly proposed as reasoning engines that triage alerts and, in autonomous deployments, issue commands t...
G. Koutidis, N. Kekatos, T. Nianios et al.· Computer Science Symposium i...· 3 citations
Distributed Energy Resource (DER) environments rely on network communication protocols to coordinate control commands, measurements, and device states across edge assets and cloud systems. Edge anomaly detection systems (ADS) monitor this traffic to identify deviations from normal communication behavior, flagging suspi...
D. Popoola, S. Bhattacharya, M. Govindarasu· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.