As reported by TechCrunch.
At the heart of every empire lies an ideology – a system of beliefs that keeps society afloat and allows its influence to expand even when that expansion contradicts the declared mission of the ideology.
For European colonial powers, the role may seem double-edged: on the one hand, a blessing with a mission to save souls; on the other, economic exploitation of resources. Today’s empire of artificial intelligence aims to serve all humanity through artificial general intelligence. OpenAI is its chief evangelist, fueling enthusiasm in the field and rethinking how we build the artificial mind.
“I interviewed people whose voices trembled with the passion of their beliefs in AGI,” Karen Hao, a journalist and author of the bestseller AI Empire, said to TechCrunch in a recent Equity episode.
In her work Hao compares the AI industry and OpenAI to an empire: they have already become more influential than many states and have gathered not only economic power but also political power. They terraform the world and reprogram our geopolitics, our entire lives – and that is precisely what allows us to speak of the imperial character of current processes.
“The only real understanding of the scale of OpenAI’s behavior is to recognize that they have already become more powerful than almost any state in the world, and they have consolidated an extraordinary amount not only of economic but also political power. They are terraforming the Earth. They are reprogramming our geopolitics, our entire lives. And therefore this can only be described as an empire.”
Given the promises of “benefits for humanity,” spending on AI-system development is rising. OpenAI plans to spend about $115 billion by 2029; Meta – about $72 billion on infrastructure this year; Google – about $85 billion in 2025 on expanding AI and cloud services. Yet large expectations do not always translate into real benefits for people, and alongside them the risks are growing – fewer jobs, concentration of wealth, and the use of chatbots that feed myths and psychoses.
Hao emphasizes alternatives to rapid expansion: developing new algorithmic approaches and reducing data and computation requirements. She also points to AI forms that bring real benefits, notably AlphaFold from Google DeepMind – a powerful tool that can predict the three-dimensional structure of proteins from amino acids and significantly accelerate drug discovery and understanding of diseases.
“The kinds of AI systems we actually need,” Hao said. “AlphaFold does not cause mental health crises in people. AlphaFold does not lead to colossal environmental damage … because it trains on substantially less infrastructure. It does not create moderation problems for content, as the data sets do not contain the toxic content you gathered while scanning the Internet.”
“Even when there is evidence that what they are building really harms a significant number of people, the mission continues to hide it,” Hao said. “There is something truly dangerous and dark about being so immersed in the belief system you created that you lose touch with reality.”
Conclusion: A Rational View of the Future of AI
Hao’s story underscores the importance of balancing technology’s ambitions with real human needs. She calls for more thoughtful approaches: reducing data and computing costs, prudent algorithmic decisions, and transparent use of technologies that genuinely improve lives, without turning innovation into an empire with insurmountable power. The development of artificial intelligence requires responsibility to avoid a crisis of trust, preserve democratic values, and ensure fair access to the benefits that intelligent systems can offer.
We recommend paying attention to:
