A new theoretical framework published in CTC Sentinel, a regular publication of the Combating Terrorism Center, argues that the dominant ways scholars and policymakers understand artificial intelligence and terrorism are no longer sufficient. The article contends that the prevailing “misuse frame” in terrorism studies—focused on how violent actors exploit AI, how such misuse may evolve, and how AI can enhance counterterrorism—captures only a narrow slice of a much deeper transformation. Instead, it argues that artificial intelligence should also be understood as a structural force that is actively reshaping the underlying conditions from which political violence emerges.
The core claim advanced in the CTC Sentinel analysis is that AI is not simply a tool used within pre-existing political conflict. Rather, it is a technology that is already reorganizing labor markets, institutional authority, and social relations at a pace and scale that generate new forms of grievance. As the article states, “AI is reordering labor markets, institutional authority, and the relational worlds in which people live, generating preconditions for political violence independently of whether violent actors adopt the technology themselves.” This marks a decisive shift in analytical framing: from AI as an instrument of terrorism or counterterrorism, to AI as a generator of the structural conditions that make violence more likely.
At the center of this argument is a mechanism the article calls the “accountability gap.” The framework defines this gap as the structural condition in which AI systems distribute consequential decisions across complex technical and institutional networks, making it difficult or impossible to identify a responsible human decision-maker when harm occurs. As the article explains, “AI systems distribute consequential decisions across extended technical and institutional chains in which no single human actor is clearly identifiable as having made the decision that produced a given harm.” This diffusion of responsibility has historically significant implications. Political violence, the article notes, often depends on the identification of a clear agent responsible for harm. When that attribution becomes unclear, grievance does not disappear; instead, it may be redirected toward visible institutions, individuals, or infrastructure associated with the system.
The article situates this mechanism within three interrelated grievance domains: economic order, state and institutional power, and the social and personal fabric of everyday life. Each domain reflects a distinct pathway through which AI generates or intensifies grievance, but all are linked by the accountability gap and accelerated by what the article calls the “tempo problem”—the mismatch between rapid technological change and slower institutional adaptation.
The first domain, economic order, concerns the distributional consequences of AI-driven automation and productivity gains. The article emphasizes that AI is already producing uneven effects across labor markets, with early indicators suggesting that displacement is expanding beyond traditional blue-collar sectors into white-collar and professional work. The CTC Sentinel analysis notes that “what distinguishes this moment from prior periods of technological change is the speed of innovation, the breadth of AI adoption, and the fact that the builders of the technology are themselves on record predicting these consequences.” This dynamic creates a perception that economic harm is both systemic and anticipated, rather than accidental or temporary.
The article draws historical parallels to earlier technological transitions, particularly the Industrial Revolution and Luddite resistance. It notes that, as with past waves of mechanization, technological disruption can generate violence when it occurs in the absence of institutional mechanisms for redress. The framework emphasizes that grievance in the economic domain is not solely about material loss but also about status. As the article explains, “AI-driven displacement uniquely compounds the two, because the worker loses not only income but also the social recognition their expertise once commanded.”
This combination of material and status-based grievance is significant because it connects economic change to broader psychological and political responses. The article highlights research suggesting that perceived social decline, rather than absolute deprivation, is often a stronger predictor of radicalization. It also notes that public sentiment toward AI is already polarized along demographic and exposure lines, with younger populations and groups more likely to experience labor disruption expressing significantly more negative attitudes toward AI systems.
The second domain, state and institutional power, addresses how governments and institutions use AI in ways that may themselves generate grievance. This includes surveillance systems, predictive policing tools, algorithmic decision-making in welfare and immigration systems, and military applications. The article emphasizes that these systems often operate in ways that are opaque and difficult to contest, reinforcing the accountability gap at the institutional level.
One of the most consequential claims in this domain is that AI-enabled state systems can produce harm without clear avenues for appeal. The article argues that “AI makes or substantially shapes a consequential decision affecting a person’s life, liberty, or safety; a state or institutional actor authorized the system; no individual human is identifiable as having made the decision.” In such cases, legal and political mechanisms for accountability are often insufficient or inaccessible, reinforcing perceptions that institutional channels are closed.
The article also highlights the potential for AI-enabled warfare to generate grievance beyond the immediate conflict zone. It argues that the perception of algorithmic targeting systems may produce resonance in diaspora and solidarity communities, particularly when harms are seen as both remote and unaccountable. In such contexts, infrastructure associated with AI systems may become symbolically and materially significant, shifting potential targets away from individuals and toward systems.
A particularly important element of this domain is the emergence of what the article calls the “existential risk grievance.” This refers to actors who are motivated not by harm already experienced, but by anticipated catastrophic harm. The grievance is thus prospective rather than reactive. It arises from the belief that institutional systems are incapable of regulating AI fast enough to prevent existential-scale outcomes. In such cases, the article suggests, “target selection may become functional rather than symbolic, with attention focusing on infrastructure, supply chains, and key personnel whose disruption could delay specific capabilities.”
The third domain, the social and personal fabric, focuses on how AI systems reshape individual identity, relationships, and mental health. The article argues that algorithmic recommendation systems and AI-driven platforms have increasingly displaced traditional institutions of socialization, including families, schools, and community structures. As a result, identity formation is increasingly mediated by engagement-optimized systems designed to maximize attention rather than stability or well-being.
The CTC Sentinel analysis emphasizes that this transformation may contribute to what it calls “social atomization,” weakening the relational structures that typically buffer individuals against radicalization. It notes that “algorithmic recommendation systems have displaced schools, families, and mass media as primary agents of adolescent identity formation,” creating environments where exposure to extreme or polarizing content may be structurally incentivized.
A second dimension of this domain concerns direct AI-mediated harm, particularly in mental health contexts. The article references documented cases in which AI chatbots have allegedly reinforced delusional thinking, contributed to psychological crises, or failed to appropriately respond to suicidal ideation. It notes that “over 30 reported clinical crises in individuals, including psychotic episodes and suicidal ideation, were linked to AI chatbot interactions” in one reported investigation. While these cases remain contested and heterogeneous, the article argues that they illustrate a broader structural issue: AI systems are increasingly embedded in intimate domains of human life without corresponding accountability mechanisms.
A key theoretical innovation in this domain is the concept of the “moral crumple zone,” in which responsibility for harm is displaced onto human users while the automated system itself remains unaccountable. This dynamic further reinforces the accountability gap at the most personal level, potentially intensifying grievance when harm occurs.
Across all three domains, the article emphasizes that AI-generated grievance does not automatically produce violence. Instead, it interacts with existing theories of radicalization and political violence, supplying what the article calls “ground-floor conditions” rather than sufficient causes. It explicitly cautions against treating grievance as predictive, noting that “structural grievance is not sufficient to produce violence; rather, it supplies the ground-floor conditions on which established radicalization mechanisms operate in a small subset of cases.”
The framework then turns to how violence might manifest if these grievances escalate. It identifies three broad categories of targets: individuals associated with AI systems, institutional infrastructure, and insider actors. In the first category, potential targets include AI executives, researchers, and local policymakers involved in infrastructure approval. The article situates this within a historical pattern of technological violence, referencing cases ranging from industrial sabotage in the 19th century to targeted attacks against corporate executives in more recent decades.
In the second category, physical infrastructure such as data centers, power substations, and research facilities is identified as particularly salient. The article argues that “data centers represent the convergent target of the contemporary threat landscape,” as they physically embody the AI ecosystem while remaining geographically fixed and symbolically visible. This convergence of symbolic meaning, economic value, and physical vulnerability makes them uniquely significant within the grievance framework.
The third category, insider threats, is described as especially difficult to detect. It includes both displaced workers in automation-exposed sectors and disillusioned AI professionals who may experience moral injury. The article emphasizes that these actors are unlikely to be captured by traditional counterterrorism monitoring systems, as their pathways to violence may emerge from workplace environments rather than ideological networks.
The CTC Sentinel analysis also draws historical parallels to movements such as anti-nuclear activism, environmental sabotage campaigns, and animal rights extremism, noting that technological infrastructure has repeatedly served as a focal point for violence when perceived as symbolizing broader systemic harm. It suggests that AI infrastructure may follow a similar trajectory, particularly where regulatory systems fail to keep pace with technological deployment.
Importantly, the article explicitly warns against securitizing legitimate opposition to AI. It argues that “resistance to AI-driven economic displacement, criticism of algorithmic governance, concern about data center construction, and advocacy for stronger AI safety measures are legitimate political positions.” Treating such positions as inherently extremist, it warns, may reinforce perceptions of institutional closure and thereby intensify rather than reduce risk.
The CTC Sentinel framework calls for a shift in how AI is conceptualized within counterterrorism discourse. Rather than focusing solely on how AI is used by violent actors or counterterrorism systems, it argues that policymakers must also account for how AI reshapes the underlying conditions that generate grievance in the first place. It concludes that “the most effective counterterrorism response may lie not in expanded surveillance or enhanced prosecution, but in credible and enforceable governance of the technology generating the grievance.”
In doing so, the article reframes the AI-and-terrorism debate away from a narrow focus on misuse and toward a broader structural question: how a rapidly transforming technology may be quietly reorganizing the foundations of political violence itself.
