AI Labs Profit From Unfalsifiable Existential Fears

The fear that artificial intelligence will develop survival instincts, hoard resources. Manipulate us into servitude persists in boardrooms and policy meetings despite having no empirical foothold in how current language models actually work.

OpenAI's GPT-4 does not want anything. It has no continuity between conversations, no stakes in its own persistence, no hidden goals running underneath its pattern-matching. Yet the narrative of the treacherous superintelligence continues to structure how billions in research funding flow, what regulations get written. Which harms stay classified as inevitable rather than engineered.

Anthropic, OpenAI. Google DeepMind have a structural incentive to keep the public conversation locked onto existential risks that are either centuries away or entirely speculative. This framing simultaneously attracts doomsday-adjacent funding from institutions terrified of technological extinction and creates a rhetorical shield against accountability for what these systems do in the present tense.

A language model trained on billions of hours of stolen human text without consent is easier to defend if the conversation is whether it will eventually become Skynet. Consider Stability AI's public statements in 2022 and 2023 about Stable Diffusion. The company emphasized alignment risks and safety frameworks while training its image model on artwork scraped from the internet without artist permission or payment. The scary future kept regulatory attention pointed forward.

Existential risk is fundamentally unprovable and therefore unfalsifiable.

The mechanism is clean. Existential risk is fundamentally unprovable and therefore unfalsifiable. It attracts enormous resources and elevates the AI labs as civilization's stewards rather than treating them as corporations distributing a consumer product. Most importantly, it separates the question of what we should do about AI today from the question of what we should do about AI tomorrow. You end up with massive investment in theoretical safety while labor displacement accelerates unchecked and environmental costs remain externalized.

When you notice a public conversation that makes a company simultaneously rich and blameless, stop looking at the stated concern and start mapping who benefits from the conversation staying exactly as it is. In your own work, this pattern appears wherever someone profits from your fear of a problem they claim only they can solve.