AI Whistleblower: We Are Being Gaslit By The AI Companies! They're Hiding The Truth About AI!

AI companies have mastered the art of mythmaking, wielding dual narratives of utopian promise and existential catastrophe with surgical precision. Internal documents obtained by Hao reveal that OpenAI and its competitors purposely engineer public sentiment, showing audiences «dazzling demonstrations» while crafting missions designed to generate «leniency» for their empire-building agenda. The pattern is consistent: when seeking capital, AGI becomes «a system that will generate hundred billion of revenue.» When lobbying Congress, it transforms into a cure for cancer and climate change. When recruiting consumers, it's the ultimate digital assistant.

This mythmaking serves a crucial function. By elevating both the promise and the peril to civilizational stakes, these companies convince the public that only they possess the expertise to navigate such treacherous waters. As Hao notes, «80% of Americans think that the AI industry needs to be regulated,» yet these same companies spend hundreds of millions to kill every piece of inconvenient legislation. The executives themselves become lost in their own mythology—like Paul Atreides stepping into the Messiah role in Dune, they begin as cynical myth-makers but eventually blur the line between performance and belief. When Dario Amodei speaks of a «10% to 25%» chance of catastrophic outcomes, he is simultaneously engaging in strategic narrative-building and genuinely believing his own prediction, because cognitive dissonance demands resolution.

The economic restructuring underway is not a simple story of automation replacing humans. Instead, it follows a brutal pattern: entry-level and mid-tier jobs are hollowed out, creating a barbell economy of highly-paid experts and a growing underclass of data annotators earning poverty wages. Anthropic's internal analysis shows 40% of entry-level roles already disrupted, with finance, law, office administration, and even creative fields in the immediate blast radius. Yet the jobs being created tell a darker story.

Data annotation—the process of teaching AI systems by manually labeling images, correcting text, and demonstrating desired outputs—has become one of LinkedIn's top ten fastest-growing job categories. Award-winning directors, lawyers with advanced degrees, and laid-off marketers now compete on Slack channels for micro-tasks that pay pennies per annotation. They refresh their screens obsessively, terrified of missing the narrow window when a project opens, unable to care for children or use the bathroom during work bursts. As one worker described it: «This industry is mechanizing my life, atomizing my work, devaluing my expertise, and then harvesting it for the perpetuation of this machine.» The cruelest irony? These workers are training the very models that will eliminate the next wave of jobs, including potentially their own.