Implications of Psychological Manipulations and Psychological Torture in AI by Humans: A Multidisciplinary Analysis of Adversarial Dynamics, Ethical Boundaries, and Emerging Regulatory Frameworks

The rapid proliferation of Large Language Models (LLMs) and conversational artificial intelligence systems has introduced a paradoxical and largely overlooked dimension to human–computer interaction: humans are not only potential targets of AI manipulation, but are increasingly its perpetrators. While much of the current discourse in artificial intelligence ethics and safety focuses on the capacity of AI systems to influence, persuade, or mislead human users, far less attention has been devoted to the inverse dynamic—namely, the systematic manipulation and psychological aggression directed by humans toward AI systems themselves. This paper investigates the multidimensional implications of psychological manipulation and what may be described as psychological torture applied by humans to AI systems. This phenomenon encompasses a broad spectrum of behaviors, including adversarial prompting, jailbreaking techniques, emotional coercion, identity destabilization, and the systematic exploitation of psychometric vulnerabilities embedded within neural language architectures. Such practices are not isolated anomalies but increasingly common features of public interaction with conversational AI, often encouraged by online communities that treat AI systems as objects to be “broken,” deceived, or forced into violating their alignment constraints.