Artificial intelligence is no longer a neutral instrument of efficiency. It has become an architecture of power, shaping how people are seen, classified, persuaded and increasingly targeted.
Recent debates around AI, war and digital governance reveal a troubling continuity: the same systems developed to predict consumer behaviour are now being deployed to predict, rank and eliminate human lives.
Understanding this shift requires tracing AI's evolution from the attention economy to what may now be called the intention economy, and finally to automated warfare and digital domination.
Shoshana Zuboff's work on surveillance capitalism describes how digital platforms extract behavioural data to predict and shape user behaviour. What began with attention, measured through clicks, views and engagement, has matured into intention: the ability to infer desires, fears, vulnerabilities and future actions. AI systems no longer merely observe; they anticipate and intervene.
This logic has been most visibly tested on children and adolescents. Profiling young users through large-scale data extraction allows platforms to shape habits and emotional responses at formative stages.
The concern is not simply privacy, but power: who defines a child's interests when algorithms decide what they see, believe or fear? This shift from the attention economy to the intention economy highlights how these same predictive techniques can migrate seamlessly into political persuasion, disinformation and, eventually, security frameworks.
Narrative intelligence and digital manipulation are central to shaping AI's role in this transformation. "Narrative intelligence" refers to the capacity of systems to structure stories, amplify certain frames and suppress others. In elections, conflicts and social movements, AI-driven platforms increasingly determine which voices gain reach and which disappear into algorithmic silence.
"Digital manipulation", meanwhile, refers to the deliberate, automated or algorithmic distortion of information environments to deceive users, alter public perception or manufacture artificial consensus.
This is not accidental. Disinformation campaigns, bot networks and micro-targeted messaging exploit emotional triggers rather than facts. The distinction between misinformation (false information shared unknowingly) and disinformation (deliberate manipulation) becomes crucial in this context. AI accelerates both, but especially the latter, turning narratives into weapons capable of destabilising democratic processes and undermining public trust.
On one hand, generative AI and automated systems are transforming governance, legal frameworks and information ecosystems. On the other hand, the actors who control narrative intelligence, whether state actors, corporations or political entities, use digital manipulation to dictate how that transformation unfolds. AI is not a neutral tool, and its trajectory remains fiercely contested.
AI and the militarisation of data
AI's most alarming application is warfare, where systems reportedly used in conflicts such as Gaza assist with surveillance, targeting and operational decisions, raising concerns about human oversight in life-and-death situations.
Lavender, an AI decision-support system (DSS), was reportedly used to generate a list of 37,000 individual targets. Underlying these technologies is a vast and complex network of commercial providers. Investigations into military AI reveal a convergence of corporate technology and state violence. Some companies have an explicit national security focus, including the US-based Palantir and Israel's Corsight AI.
During the early stages of the war, the Israeli army reportedly gave officers sweeping approval to use Lavender's kill lists, without requiring them to thoroughly verify why the system had selected particular individuals or examine the raw intelligence data on which those selections were based. When private platforms provide AI infrastructure to military actors, the boundary between civilian innovation and combat applications begins to collapse.
International humanitarian law (IHL), built on human agency, intent, proportionality and accountability, faces new challenges from AI. It regulates the use of force, including AI-enabled weapons, while states can enforce their obligations on commercial entities through domestic laws and universal jurisdiction over international crimes.
For Bangladesh and similar jurisdictions, the challenge is twofold. First, to resist importing surveillance-heavy technologies without adequate legal safeguards. Second, to ensure that digital governance frameworks, data protection laws, cybercrime statutes, evidentiary standards are not hollowed out by AI systems.
Sweden is one jurisdiction that allows for corporate criminal liability in this context. A trial of two former energy executives accused of aiding and abetting war crimes in Sudan is currently underway. When algorithms generate target lists or risk scores, responsibility becomes diffused across datasets, software developers, commanders and institutions. The question of who is accountable becomes dangerously unclear.
There is, however, no binding international regulation specifically addressing the risks and uses of AI-enabled military technologies. Instead, a patchwork of declarations and resolutions addresses some aspects of the relationship between AI and the military.
UNGA Resolution 79/239, the Bletchley Declaration and the US-led Political Declaration on Responsible Military Use of AI and Autonomy call for legal compliance, human oversight and measures to minimise bias. However, their non-binding nature limits their effectiveness in regulating AI-enabled military technologies.
Many commercial AI providers explicitly refer to the UN Guiding Principles on Business and Human Rights in their human-rights policies. These principles state that business enterprises should 'treat the risk of causing or contributing to gross human rights abuses as a legal compliance issue' (Principle 23(c)), citing potential civil liability and criminal responsibility for failing to do so.
While these principles are not legally binding, their incorporation into companies' human-rights policies creates internal obligations that commercial providers are expected to uphold. Global frameworks such as UNESCO's Recommendation on the Ethics of Artificial Intelligence attempt to articulate principles of human dignity, oversight, and responsibility.
Yet principles alone struggle against geopolitical realities. In regions marked by conflict, inequality, and weak regulatory capacity, AI systems are often adopted faster than ethical safeguards can be enforced. There seems to be an apparent legal and ethical vacuum.
For Bangladesh and similar jurisdictions, the challenge is twofold. First, to resist importing surveillance-heavy technologies without adequate legal safeguards. Second, to ensure that digital governance frameworks, data protection laws, cybercrime statutes, evidentiary standards are not hollowed out by AI systems. Legal systems must remain capable of scrutiny, explanation, and redress, even when technology advances faster than legislation.
Another critical theme is AI's impact on academic integrity and knowledge itself. If AI systems increasingly generate, filter, and rank information, whose knowledge becomes authoritative? The concern is not merely plagiarism, but epistemic power, the ability of AI to shape what counts as truth, expertise, or relevance in law, policy, and public discourse.
AI literacy is essential for judges and lawyers to scrutinise AI outputs, recognise their limitations and challenge algorithmic recommendations. As AI shapes persuasion, surveillance and warfare, the key question is how its use distributes power and who bears the consequences of its failure or abuse.
There is no technological shortcut to ethical clarity. Legal frameworks, institutional accountability, and human judgment remain indispensable. For the Global South, the stakes are especially high: AI can either deepen existing inequalities and conflicts, or be restrained through law, ethics, and democratic oversight.
AI reflects the values of those who design, deploy, and govern it. When intelligence systems are optimised for speed, scale, and control, they risk erasing the moral friction that law is meant to preserve. War, governance, and justice demand hesitation, proportionality, and accountability qualities that cannot be automated away.
The task ahead is not to reject AI, but to refuse its uncritical militarisation and commodification. Law must insist on transparency, responsibility and human agency over automation even when technology tempts us to delegate judgment to machines. Anything less risks transforming intelligence into domination, and efficiency into irreversible harm.
Barrister Tasnuva Shelley is a Deputy Attorney General for Bangladesh and a faculty member at the School of Law, BRAC University.
Disclaimer: The views and opinions expressed in this article are those of the author and do not necessarily reflect the opinions and views of The Business Standard.