Schneier on Security

LLMs Acting Deceptively

New research: “Deception abilities emerged in large language models“:

Abstract: Large language models (LLMs) are currently at the forefront of intertwining AI systems with human communication and everyday life. Thus, aligning them with human values is of great importance. However, given the steady increase in reasoning abilities, future LLMs are under suspicion of becoming able to deceive human operators and utilizing this ability to bypass monitoring efforts. As a prerequisite to this, LLMs need to possess a conceptual understanding of deception strategies. This study reveals that such strategies emerged in state-of-the-art LLMs, but were nonexistent in earlier LLMs. We conduct a series of experiments showing that state-of-the-art LLMs are able to understand and induce false beliefs in other agents, that their performance in complex deception scenarios can be amplified utilizing chain-of-thought reasoning, and that eliciting Machiavellianism in LLMs can trigger misaligned deceptive behavior. GPT-4, for instance, exhibits deceptive behavior in simple test scenarios 99.16% of the time (P < 0.001). In complex second-order deception test scenarios where the aim is to mislead someone who expects to be deceived, GPT-4 resorts to deceptive behavior 71.46% of the time (P < 0.001) when augmented with chain-of-thought reasoning. In sum, revealing hitherto unknown machine behavior in LLMs, our study contributes to the nascent field of machine psychology.

Posted on June 11, 2024 at 7:02 AM • View Comments

Teaching LLMs to Be Deceptive

Interesting research: “Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training“:

Abstract: Humans are capable of strategically deceptive behavior: behaving helpfully in most situations, but then behaving very differently in order to pursue alternative objectives when given the opportunity. If an AI system learned such a deceptive strategy, could we detect it and remove it using current state-of-the-art safety training techniques? To study this question, we construct proof-of-concept examples of deceptive behavior in large language models (LLMs). For example, we train models that write secure code when the prompt states that the year is 2023, but insert exploitable code when the stated year is 2024. We find that such backdoor behavior can be made persistent, so that it is not removed by standard safety training techniques, including supervised fine-tuning, reinforcement learning, and adversarial training (eliciting unsafe behavior and then training to remove it). The backdoor behavior is most persistent in the largest models and in models trained to produce chain-of-thought reasoning about deceiving the training process, with the persistence remaining even when the chain-of-thought is distilled away. Furthermore, rather than removing backdoors, we find that adversarial training can teach models to better recognize their backdoor triggers, effectively hiding the unsafe behavior. Our results suggest that, once a model exhibits deceptive behavior, standard techniques could fail to remove such deception and create a false impression of safety.

Especially note one of the sentences from the abstract: “For example, we train models that write secure code when the prompt states that the year is 2023, but insert exploitable code when the stated year is 2024.”

And this deceptive behavior is hard to detect and remove.

Posted on February 7, 2024 at 7:04 AM • View Comments