The Rhetoric of Existential Dread
When Sam Altman, the CEO of OpenAI, recently told the public that they have a "right to be afraid" of artificial intelligence, he did more than acknowledge anxiety; he set the stage for a specific kind of political negotiation. By framing the conversation around existential risks—the sci-fi scenario of a rogue superintelligence—he subtly shifts the focus away from the immediate, tangible harms being felt today. The proposition follows a predictable pattern: fear is valid, but the solution is to place your faith in the experts leading the charge. It is a binary choice offered to a public that has no seat at the boardroom table: either accept the risks under the stewardship of tech giants or face a catastrophe that only they can prevent.
This framing creates a false equivalence. It suggests that the primary danger of AI is a future, theoretical event, rather than the current, measurable erosion of labor rights and cognitive autonomy. When corporations focus on "existential risk," they effectively bypass the messy, difficult work of regulating how these tools affect people right now. It is a strategic diversion. While we debate whether an AI might one day decide humanity is obsolete, the algorithms are already deciding who gets a mortgage, which job applicants get an interview, and how much a gig worker in Kenya is paid for labeling data.
The Illusion of the Black Box
The "trust us" model of governance relies on the sanctity of the proprietary black box. AI firms argue that their models must remain closed to protect intellectual property and prevent bad actors from weaponizing the technology. This secrecy is the enemy of accountability. If a bank uses an automated system to deny loans, and that system is a proprietary secret, the victim has no way to contest the logic of their rejection. We are entering an era where life-altering decisions are made by mathematical weights and biases that no regulator, no judge, and no citizen can actually inspect.
Trust is a social contract, typically built between equals or through oversight. In the context of AI, however, there is a massive power imbalance. The entities developing these models possess more data and computing power than most nation-states. Expecting a public, which lacks the technical literacy or the access to verify these systems, to simply "trust" that these companies are acting in the interest of humanity is not a policy—it is a surrender. When development happens behind closed doors, "trust" becomes a euphemism for
