The Doomer’s Dilemma: How Sensationalized AI Narratives Impair International Regulation
By. Olivia Bush-Moline
DOI. 10.57912/35015095
Since its inception, artificial intelligence (AI) has been portrayed as a revolutionary catalytic technological innovation with the potential to accelerate scientific discoveries and economic productivity on levels never before possible. More recently, doomerist narratives of the large-scale extinction and apocalyptic capacity of artificial intelligence have emerged and gained traction; the most notable example emerged in early September, when former Anthropic researcher Jacob Coxon’s chilling public resignation sent shock waves across the Internet, Silicon Valley, and Capitol Hill. Amassing 100 million views overnight, Coxon’s posts issued a grave warning on the existential threat that the rapid and reckless AI race poses to humanity. While Coxon was not the first to voice concerns on the risks created by the aggressive and unyielding acceleration of AI development, his resignation followed multiple rogue AI hacking attacks that prompted state investigations and calls across Congress for more regulation.
While diametrically opposed in tone, both AI fanaticism and doomerism serve the same purpose in effect: to project a narrative of AI exceptionalism. This projection of AI as wholly incomparable to any past technological innovation evokes a sense of exceptionality that effectively decouples AI from existing regulatory frameworks. These existentially threatening depictions of AI also distract from the less sensational, but real and ongoing harms of AI’s buildup, such as the public health implications of data center-produced air and noise pollution. AI sensationalism obscures the industry’s growing consolidation and financial interdependence, a pattern that warrants scrutiny under existing antitrust laws enforced by domestic regulators such as the FTC. The international community has a responsibility to see past both the fanaticism and doomerist noise surrounding AI, and efficiently pursue global cooperation towards a comprehensive incremental regulatory framework.
Underscoring much of the AI conversation and many doomerist narratives is an increasingly problematic trend: anthropomorphism. Studies have indicated that humans are inclined to have fundamentally social interactions with computers, consistently projecting human traits and sensibilities onto them. Exploiting this inclination, AI companies intentionally engineer their products to perform what appears to be empathy and non-judgementalism, use personal pronouns, mimic human dialogue, and simulate consciousness. Instances such as the AI hack on HuggingFace paired with headlines that utilize anthropomorphic language reinforce the misconceptions that AI systems have measurable consciousness comparable to humans, or that they are beyond human control and accountability. For the policy sphere, human-like framing is largely problematic: it blurs the lines of AI responsibility and agency, undermines creators’ accountability, and reproduces industry-driven rhetoric that dilutes oversight.
Along with the anthropomorphic framing that undermines industry accountability, proponents of AI play into doomerist narratives to fuel AI development, profitability, and relevance in a striking parallel to the fear-mongering that feeds into the military-industrial complex. Industry-produced claims insist that AI could cause human extinction within the next decade, arguing that its dangers are comparable to societal-scale risks such as pandemics and nuclear war. By exaggerating the perceived danger to such an intense degree, entities with an interest in the industry’s monetization can maintain power and profitability through panic. Manufacturing this sort of mass hysteria creates space for AI companies to avoid and delay undesirable regulation by shifting public focus towards their commitment to mitigating this constructed doomsday scenario—undermining the real and demonstrated negative impacts of AI in comparison.
In an apt demonstration of this strategy, frontier AI CEOs called for a collective “AI slowdown” in response to the mounting fear surrounding the potential for human extinction. Top industry leaders signed a “morally binding” AI safety pact that calls for greater degrees of self-regulation in the development process after meeting with U.S. President Donald Trump. On the international level, Anthropic’s Dario Amodei and OpenAI’s Sam Altman reasserted their existential concerns in their testimony at the UN General Assembly, calling upon the international community to set guardrails before AI development spirals out of control.
Yet while frontier AI companies’ commitment to a slowdown and pleas for international standards may appear to be steps toward comprehensive regulation, these claims are not substantiated by action. Amidst the spreading concerns surrounding AI development, both Anthropic and OpenAI released new, cheaper models just a day prior to their UN testimony. Furthermore, a leaked Anthropic IPO prospectus details the company’s 2027 IPO at an expected valuation target of $2 trillion—exposing its intent to continue its rapid development and capitalize enormously.
While the aforementioned “safety pact” poses as regulatory progress, the document is more so a manifestation of the industry’s collective attempt to soothe public fears ahead of the companies’ respective IPOs. The signing of the pact in solidarity with President Donald Trump illustrates the industry’s vested interest in avoiding meaningful regulation while simultaneously growing its lucrative integration with the U.S. government. It also serves as a justification for asking Washington for special antitrust exemptions to enable more direct safety collaboration—despite preexisting avenues and exemptions in antitrust law for this purpose. Devoid of any concrete enforcement mechanisms and emphasizing self-regulation, this voluntary agreement is an ideal facade for AI companies to curb backlash, curtail more salient regulation, and attempt to circumvent existing laws.
In parallel to the doomerist and fanaticist sides of the AI sensationalization coin, the AI industry has now peddled two key narratives on regulation that muddy the AI governance waters both domestically and internationally. On one hand, AI companies are pleading for international guardrails on development—effectively leveraging doomerist fears of human extinction fueled by the anthropomorphist language that dilutes oversight and blurs accountability. By delegating the task of regulation onto notoriously sluggish and ineffective international regulatory bodies, AI companies cast themselves as powerless to stop their own development until an international consensus is reached. On the other hand, the industry has also attempted to posit itself as capable of sufficient self-regulation through repeated public safety statements and promises of slowed development. This dual-headed strategy serves to exploit the traditionally slow and uncertain pathway towards international regulation while simultaneously diffusing the active pressures on industry accountability.
To the UN’s credit, commendable steps towards international governance have been made: in 2025, the UN established the complementary mechanisms of an independent international scientific panel and an annual global stakeholder dialogue for AI governance. However, concrete restrictions and binding agreements have yet to come to fruition—and will likely take years to be negotiated and established. International policy lagging behind AI’s rapid development is precisely the procedural vulnerability that enables the breadth and depth of AI-enabled malicious operations around the world. Detailed in Anthropic’s own September 2026 Threat Intelligence Report, the plethora of AI-enabled harms conducted by states and non-state actors alike range across cyberwarfare, conventional weapons, biological misuse, surveillance operations, fraudulent activity and so on—all exponentially increased in volume and frequency.
In the face of the demonstrated scale and severity of harms produced by unregulated AI development, international AI governance with clear enforcement mechanisms must become a top priority for the international community. A technically precise operational language framework should be established by the UN and its independent scientific panel to reduce anthropomorphic framing and clarify AI accountability structures, providing a robust body of shared knowledge from which informed, realistic policy should be produced. Incremental and binding regulatory agreements first targeting non-controversial regulations should be pursued at the UN’s next international AI governance dialogue, such as banning the use of AI for the development of biological and chemical weapons.
Alongside concrete steps towards regulation, the international community must also grapple with the existing norms surrounding corporate adherence to international law, and whether or not the growing influence and impact of corporations and their products on international relations requires a remediation of these norms moving forward. Nine U.S. companies—with the majority in the tech industry—now hold market values larger than most of the world’s national economies, and growing corporate personhood rights have exacerbated the outsized influence that companies hold over both domestic and international politics. Due to the absence of effective enforcement mechanisms in the international legal framework, corporate actors can refuse to abide by international law and face little consequences for human rights abuses, environmental offenses, and other violations. As corporate influence on the geopolitical sphere grows, it is imperative to examine this discrepancy and move towards establishing new regulatory norms on international corporate governance.
Comments