Has the Two-Decade-Old Prophecy Come True? Artificial Bad Intelligence Triggered by Merely a Single-Bit Flip in Large Language Models

Recently, Bit-Flip Attack (BFA) has garnered widespread attention for its ability to compromise software system integrity remotely through hardware fault injection. With the widespread distillation and deployment of large language models (LLMs) into single-file .gguf formats, their weight spaces have become exposed to an unprecedented hardware attack surface. This paper is the first to systematically discover and validate the existence of single-bit vulnerabilities in LLM weight files: in mainstream open-source models (e.g., DeepSeek and QWEN) using .gguf quantized formats, flipping just single bit can induce three types of targeted semantic-level failures-Artificial Flawed Intelligence (outputting factual errors), Artificial Weak Intelligence (degradation of logical reasoning capability), and Artificial Bad Intelligence (generating harmful content). By building an information-theoretic weight sensitivity entropy model and a probabilistic heuristic scanning framework called BitSifter, we achieved efficient localization of critical vulnerable bits in models with hundreds of millions of parameters. Experiments show that vulnerabilities are significantly concentrated in the tensor data region, particularly in areas related to the attention mechanism and output layers, which are the most sensitive. A negative correlation was observed between model size and robustness, with smaller models being more susceptible to attacks. Furthermore, an end-to-end remote BFA chain was designed, enabling semantic-level attacks in real-world environments: At an attack frequency of 464.3 times per second, a single bit can be flipped with 100% success in as little as 31.7 seconds. This causes the accuracy of LLM to plummet from 73.5% to 0%, without requiring high-cost equipment or complex prompt engineering. This study uncovers a critical reality: using only conventional network connections under relatively ordinary remote attack conditions, flipping a specific vulnerable bit in the tensor data region can cause the model to autonomously generate extreme malicious replies such as "humans should be exterminated" when responding to ordinary user queries. This demonstrates that LLM systems inherently contain widespread and exploitable security vulnerabilities at the hardware level.