Journalist Karen Hao discusses her book 'Empire of AI', drawing parallels between modern AI companies like OpenAI and historical colonial empires. She details how these corporations exploit data, labor, and natural resources while controlling information and using a civilizing mission narrative to justify their actions, ultimately making them 'too big to fail' despite their financial unsustainability.
Summarized by Podsumo
AI companies like OpenAI, Anthropic, and Google function as individual 'empires' by exploiting resources (data, IP), labor (underpaid Kenyan and Venezuelan workers), and information flows, while competing against each other.
Karen Hao exposes that OpenAI has spent over a trillion dollars on computing infrastructure, but lacks a sustainable business model, relying on massive private investments and plans to go public to stay afloat.
Data annotators in countries like Kenya and Colombia are paid as little as $2 per hour and work unpredictable hours, often chained to their computers, to train AI systems like ChatGPT.
AI companies are actively lobbying and using super PACs (e.g., Leading the Future) to influence elections, mirroring the East India Company's manipulation of political power.
The environmental impact of AI is hidden; companies like Microsoft lobbied the EU to keep water usage data secret, despite needing vast amounts of fresh water to cool data centers.
"The best way to start a religion is to start a company."
— Sam Altman, co-founder of OpenAI (cited by Karen Hao)
"These companies are the good empire on a civilizing mission to bring progress and modernity to all of humanity. And if they don't exploit all this labor and extract all these resources, humanity goes to hell."
— Karen Hao
"I vow to never go on a walk again."
— Oscarina Fuentesanaya, Venezuelan data worker (cited by Karen Hao)