Nathan Shipper sits in his office reviewing a $13,000 monthly bill for personal AI software. Shipper, the CEO of Every, told Business Insider he uses OpenAI’s Codex to draft his responses and manage his scheduling, reflecting a new reality where management itself is being automated. For the corporate elite, AI is a high-priced assistant; for the state, it is becoming the very foundation of national security.

On May 29, OpenAI launched 'Rosalind Biodefense,' a program providing federal agencies free access to a specialized life-sciences model. According to the company’s statement, GPT-Rosalind outperforms previous models in chemistry and experiment design, and will be used by the U.S. government for pandemic preparedness and outbreak-response planning. The move marks the first time a private lab has offered such a specialized model to the state at this scale.

Read together, these developments describe the Cognitive Enclosure: a world where the state leases its defense and epidemiological modeling from private monopolies. While the CEO pays $13,000 to automate his inbox, the public is told to trust proprietary algorithms with the security of the biological commons. The threat is not just the virus, but the privatization of the tools used to fight it. As OpenAI and Anthropic gate their most powerful models, they establish a security rent system that the hollowed-out public sector can no longer afford to ignore.