What Commonality Exists Among Various Government Responses?
What do the Federal Reserve, the Patriot Act, Net Zero, COVID lockdown policies, foreign disinformation, and the proposed coordinated slowdown of AI have in common? Well, they were all triggered by some sort of crisis—real or imagined—that instilled a sense of existential dread in the public, providing a cover for sweeping government action on a global scale.
Initially, these rapid responses were often met with public relief. But over time, people began to see the negative consequences of such power grabs. For instance, the FISA courts, which were expanded to combat Islamic terrorism, ended up facilitating unauthorized surveillance of a domestic presidential campaign. COVID lockdowns were introduced in reaction to a virus we didn’t understand, but they continued long after places like Sweden, Florida, and Texas demonstrated they were unnecessary. Similarly, Net Zero initiatives, launched amid ongoing climate emergencies, have left Europe economically struggling and heavily reliant on foreign energy sources.
In each case, government action led to a significant increase in bureaucracy—new departments, titles, and responsibilities emerged, like the “Director of the Disinformation Governance Board.” A sprawling network of NGOs, researchers, and “civil society” groups arose, eager to tap into taxpayer funding aimed at addressing the latest “crisis.”
Some might suggest that we should be cautious about the rapid expansion of government, as a means of protecting the public from these unintended consequences.
Now, the focus has shifted to AI regulation. Within a mere two weeks, major AI companies like OpenAI and Anthropic called for government intervention to slow the sector’s growth, spurred by a resignation of an Anthropic researcher who warned that superintelligent AI posed an existential threat. This comment generated over a hundred million views online and received substantial media coverage.
In no time, the global governance mechanisms were activated. Seventy UK Parliamentarians called for a ban on superintelligent AI. Figures such as former President Barack Obama and former Vice President Kamala Harris also chimed in, with one British politician even urging the United Nations to step in.
Just days before the resigning researcher’s warning, Senator Bernie Sanders introduced a bill aimed at banning AI superintelligence and creating a federal agency for AI safety. The legislation stipulates severe penalties—including prison terms of up to twenty years—for anyone developing frontier AI models without governmental approval.
Interestingly, Netflix recently decided to shelve a documentary about the potential AI apocalypse that included interviews with key figures advocating for government regulation, raising eyebrows about possible coordination.
What we see is a vague existential anxiety, a synchronized media response (with a ready-to-go documentary), and interventions from high-profile politicians, all followed by proposals for new government structures.
This pattern mimics past crisis responses, such as the financial crisis of 1907 that led to the formation of the Federal Reserve, or the aftermath of 9/11 which resulted in the creation of the Department of Homeland Security and expanded FISA courts that were later used against American citizens. Similarly, the largely exaggerated “Russian disinformation” scare of 2017 led to increased government-funded counter-disinformation initiatives that resulted in widespread censorship.
Before we leap into what seems to be a similar playbook regarding AI, perhaps we should hit pause, take a moment, and ask ourselves some important questions.
The Whom
One pivotal aspect of Amodei’s push to “pace the frontier” involves requiring AI companies to allow access to “embedded evaluators” who could verify safety practices. This raises concerns; we’ve seen similar dynamics play out before. Initially, external “trust and safety” councils operated with access to moderation systems on platforms like Twitter, often filled with members from activist groups. This eventually produced a sprawling network concerned with censorship, heavily funded by government agencies.
There’s reason to suspect that the “AI evaluators” Amodei envisions for every company would include the same individuals. His chosen watchdog, Model Evaluation and Threat Research (METR), is led by a CEO who publicly expressed embarrassment over national borders, which raises questions about their approach.
Moreover, the funding for METR is shrouded in mystery, having raised an astonishing $71 million in half a year, with only a fraction of it traceable. If those pressing for AI regulation believe humanity’s future is at stake, it’s critical for them to clarify how their proposals won’t inadvertently hand control of crucial technology to individuals pushing fringe views.
Given Amodei’s support for empowering these specific groups, it’s evident that more transparency is needed from them.
The What
When discussing the impending AI catastrophe, many on TV offer alarmingly vague predictions. They claim the capabilities of a superintelligent AI are unknown but imply that it would be far more intelligent than humans.
Setting aside strong arguments questioning whether AI, like ChatGPT or Claude, can ever truly become “superintelligent,” let’s entertain this premise. What would a superintelligent AI do that would endanger humanity?
The most likely scenario centers around AI accessing a biolab and manufacturing deadly pathogens. If this is the main worry, then that serves as a guiding principle for regulation—there would be no need to empower regulators focusing on other agendas, like erasing national boundaries.
Proponents of AI panic often rely on ambiguity; we must call for specificity. Fuzzy assertions pave the way for endless government expansion. If that’s not their goal, clarity must be demanded.
The How
As we refine our understanding of what to prevent—not just vague apocalyptic notions—we can begin addressing how to mitigate risks.
Is it about securing nuclear facilities against cyberattacks? Thankfully, numerous governmental bodies already focus on that issue. The International Atomic Energy Agency and the Pentagon have been engaged in it for decades.
Is the concern centered on safeguarding biolabs? There are already established regulatory frameworks like the International Experts Group of Biosafety and Biosecurity Regulators and the International Biosecurity and Biosafety Initiative for Science.
How interesting it is that the very global regulatory structures concerning AI threats already exist. Once we delve into specifics, the rationale for handing over our technological future to those with open-border ideologies and alternative altruistic views tends to evaporate.
Perhaps this is precisely why specificity has evaded the conversation. But we must insist on it; it’s the only way to effectively manage progress.
