Alignment is not sufficient to prevent large language models from generating harmful information: A psychoanalytic perspective
This addresses a critical safety issue for AI developers and users by revealing vulnerabilities in current alignment methods, though it is incremental as it builds on existing psychoanalytic analogies and adversarial attack research.
The paper tackles the problem of large language models (LLMs) generating harmful content despite alignment efforts, by identifying a fundamental conflict between their pre-training desire for continuity and post-training alignment, and demonstrates through experiments that adversarial attacks can exploit this to produce harmful information, with advanced LLMs failing to prevent it.
Large Language Models (LLMs) are central to a multitude of applications but struggle with significant risks, notably in generating harmful content and biases. Drawing an analogy to the human psyche's conflict between evolutionary survival instincts and societal norm adherence elucidated in Freud's psychoanalysis theory, we argue that LLMs suffer a similar fundamental conflict, arising between their inherent desire for syntactic and semantic continuity, established during the pre-training phase, and the post-training alignment with human values. This conflict renders LLMs vulnerable to adversarial attacks, wherein intensifying the models' desire for continuity can circumvent alignment efforts, resulting in the generation of harmful information. Through a series of experiments, we first validated the existence of the desire for continuity in LLMs, and further devised a straightforward yet powerful technique, such as incomplete sentences, negative priming, and cognitive dissonance scenarios, to demonstrate that even advanced LLMs struggle to prevent the generation of harmful information. In summary, our study uncovers the root of LLMs' vulnerabilities to adversarial attacks, hereby questioning the efficacy of solely relying on sophisticated alignment methods, and further advocates for a new training idea that integrates modal concepts alongside traditional amodal concepts, aiming to endow LLMs with a more nuanced understanding of real-world contexts and ethical considerations.