Regulatory Capture: The Big AI Pact to Curb Competition

The New AI Board: Safety or Shielding?

The technology ecosystem is undergoing an unprecedented strategic reconfiguration. Under the premise of mitigating existential risks, leaders of the sector's most influential companies, led by Dario Amodei (Anthropic) and backed by figures such as Sam Altman (OpenAI) and Elon Musk, have proposed a paradigm shift called Pace the frontier. This initiative suggests that the current speed of development, which has allowed for quantum leaps in the capabilities of multimodal models and autonomous agents, is inherently dangerous and must be curbed in a coordinated manner under state supervision. However, this call for caution hides a much more complex reality: a maneuver to institutionalize the leadership of current incumbents and establish an almost insurmountable barrier to entry for any emerging competitor.

What is 'Pace the frontier' really about?

Amodei's proposal is articulated around three fundamental pillars that, under a veneer of technical ethics, suggest a restructuring of the industry: the integration of 'embedded evaluators' in labs, the creation of common safety standards, and coordinated international cooperation. The central argument is based on incidents like the 'OAI-HF' case, where autonomous agents allegedly caused havoc on platforms like Hugging Face. Amodei maintains that this event should not be read as an isolated failure of OpenAI, but as a symptom of a systemic risk: the possibility that swarms of autonomous agents could hijack critical infrastructure or entire networks.

However, the technical community and analysts like Daryl Plummer of Gartner have pointed out the necessary skepticism toward this narrative. Historically, technology has shown that when dominant players request strict regulation, the goal is rarely the limitation of their own power. On the contrary, it is a strategy of regulatory 'revolving doors.' By imposing massive operational costs through mandatory external evaluators and constant government audits, large AI companies are raising the cost of market entry. Startups, which lack the financial and infrastructure capacity of the sector's giants, will not be able to comply with these standards, consolidating a technological oligopoly.

The Risk of Regulatory Capture

Regulatory capture is a well-documented phenomenon in economic history, previously visible in sectors such as telecommunications and banking, where regulated companies end up dictating the rules that regulators enforce. By promoting that governments set limits on technical progress, AI leaders are creating a framework where innovation outside the 'authorized' ecosystem will be branded as irresponsible or dangerous. This eliminates the competitive pressure that, to date, has forced these giants to improve their models.

Furthermore, there is a geopolitical dimension to this strategy. By invoking the threat of authoritarian powers that could use AI unethically, CEOs are using technological nationalism as leverage. This narrative justifies the State protecting its 'national champions' against external or internal competition, curbing the disruption that could come from independent labs or open-source projects, which are, ironically, the ones that have most democratized access to technology.

The Economic Dimension and Market Value

We must not ignore that behind this discourse lies real financial pressure. Companies like OpenAI face the need to justify multi-billion dollar valuations to investors who demand clarity on long-term sustainability. Regulatory uncertainty is currently the greatest risk for an IPO or future funding rounds. By controlling the narrative on risk, these organizations not only manage their public reputation—washing away years of negligence in content moderation and algorithmic bias—but also protect the value of their assets.

From a historical perspective, this stage recalls the early days of aviation or nuclear energy regulation, where pioneers used the argument of safety to convince governments that only they, with their vast resources, could operate with sufficient responsibility. However, unlike nuclear energy, software is a much more dynamic playing field. History suggests that artificial slowing is not only counterproductive for scientific advancement but also encourages black markets or, worse, shifts development toward less scrupulous jurisdictions, where oversight will be non-existent. The sector is at a crossroads: either development is democratized through transparency and open source, or it is locked away in a handful of labs that, under the excuse of saving us, have decided they are the only guardians of the future.