Manufactured Consent in the Age of Artificial Intelligence: Rethinking Criminal Responsibility for Psychological Manipulation

The proliferation of artificial intelligence technologies has engendered an unprecedented capacity for large-scale psychological manipulation, fundamentally challenging the adequacy of existing criminal law frameworks. Contemporary AI systems, encompassing algorithmic micro-targeting, deepfake generation, emotionally adaptive chatbots, and behavioural prediction engines possess the technical capability to engineer what Edward Herman and Noam Chomsky identified as the ‘manufacturing of consent’, now at granular, automated, and personalised scale. Where traditional criminal law predicates liability on identifiable human volition, AI-driven manipulation operates through diffuse, probabilistic architectures that obscure the causal nexus between technological design and individual psychological harm. This paper examines the doctrinal and conceptual tensions that arise when artificial intelligence is instrumentalised as a vehicle for psychological coercion, false belief induction, and consent vitiation in contexts ranging from electoral disinformation and financial fraud to coercive control and hate speech amplification. The paper critically interrogates three foundational assumptions of criminal responsibility, the actus reus of manipulation, the mens rea of AI principals and intermediaries, and the causal standard of psychological harm in light of the emergent technical capacities of generative and adaptive AI. It argues that the dominant frameworks of intermediary liability, including Section 79 of the Information Technology Act, 2000, and analogous safe harbour doctrines in comparative jurisdictions, are constitutively inadequate to accommodate algorithmic manipulation as a discrete category of criminally cognisable conduct. Drawing upon the doctrine of innocent agency, the concept of wilful blindness, and comparative analysis of the European Union’s Artificial Intelligence Act, 2024, the paper proposes a tripartite liability model predicated on design culpability, deployment intent, and systemic risk threshold. The paper concludes that the criminal law must reconceptualise consent, causation, and agency in a manner cognisant of the non-linear, emergent, and probabilistic nature of AI-mediated harm, and that the absence of a bespoke legislative framework for AI-driven psychological manipulation constitutes a critical lacuna in India’s digital justice architecture.