When journalist and author Karen Hao talks about artificial intelligence, people listen. Hao, known for her deep, investigative coverage of the industry, is now setting her sights on one of the most influential and contentious companies in the history of tech: OpenAI.
Once a small nonprofit that promised to build artificial intelligence for the good of humanity, OpenAI is now a megacorporation worth nearly $90 billion.
Her new book, Empire of AI: Dreams and Nightmares in Sam Altman’s OpenAI, along with her recent interviews, offers a colorful picture of how this empire was created — its successes, contradictions, and the profound implications of its very existence.
From Transparency to Secrecy
Its early leaders, including Elon Musk and Sam Altman, declared a lofty mission when they founded the organization in 2015: to push for the development of advanced AI in such a way that guaranteed the technology would serve people rather than a small elite.
The organization vowed to be radically transparent and said it would share its research openly with the world.
However, that culture of openness evaporated soon after, Hao said. As competition heated up and the potential for large profits became apparent, OpenAI moved toward secrecy:
- Models that used to be openly available became private.
- Research papers grew thinner.
- A capped-profit structure was introduced, allowing investors to make substantial returns while maintaining the appearance of a mission-driven nonprofit.
To Hao, this was the turning point — the moment when the ideal was eclipsed by commercial ambition.
The Three Clans Inside OpenAI
Hao’s reporting shows that OpenAI was never a monolith. Instead, it was marked by three internal “clans”:
- The Idealists – Believers in openness, prioritizing humanity’s collective good.
- The Pragmatists – Advocates of rapid scaling as the only way to survive against competitors.
- The Profit Seekers – Focused on translating breakthroughs into profitable products.
The friction between these groups shaped many of the company’s most contentious moves. While it fueled rapid innovation, it also created governance crises — most dramatically in November 2023, when Sam Altman was briefly removed as CEO, only to be reinstated days later after a very public tug-of-war among the board, employees, and investors.
The Machinery of Empire
Why does Hao call OpenAI an “empire”? Because, she argues, it reflects the same patterns as historical empires: the extractive use of resources, labor, and territory to dominate.
- Data: OpenAI trained its models on massive swathes of the internet — images, text, and code produced by millions of people, none of whom were compensated or gave consent. Hao views this as modern resource extraction: treating public creativity as raw material for private profit.
- Labor: Behind AI systems lies an invisible workforce of low-paid contractors in the Global South who label data and moderate harmful content. In places like Kenya, some workers earned only a few dollars an hour while being exposed to violent and disturbing images, leaving long-lasting psychological harm.
- Energy & Environment: Training enormous AI models requires staggering amounts of electricity and water. Hao notes that data centers strain local resources, especially in water-stressed regions. At the current pace, AI’s energy consumption could soon exceed that of entire nations.
The China Question
Silicon Valley executives often argue that the U.S. must accelerate AI development to compete with China. Hao disputes this narrative.
She contends it is frequently used as a distraction from domestic regulation. While China is indeed a global AI leader, it also has some of the world’s strictest regulations, second only to the European Union.
By presenting AI as a geopolitical arms race, companies like OpenAI can justify secrecy and speed, sidestepping the democratic oversight that might otherwise constrain their expansion.
Democracy in the Age of AI Empires
For Hao, the implications go beyond the tech industry. She warns that the unchecked consolidation of AI power in a handful of corporations poses a direct threat to democracy.
- Ordinary people lose agency over their futures when corporations control information, power, labor, and skills.
- Decisions about how AI is built and deployed are made in secret, driven by investor interests rather than the public good.
Hao describes this as a return to “an age of empire”, in which communities are left powerless as giant institutions extract from them without giving back.
The Human Side of the Story
One of Hao’s strengths is her focus on human stories beneath the empire.
She highlights:
- Engineers caught in ethical dilemmas.
- Workers paid to review traumatic content.
- Communities facing environmental costs.
Through these stories, readers encounter AI not as an abstract technology but as something built by people, run by people, and affecting people in profound ways.
An Invitation to Choose Differently
Despite her critical lens, Hao does not leave readers in despair. She stresses that alternatives exist:
- AI can be developed transparently, fairly, and sustainably.
- Governments, companies, and citizens can demand accountability, regulate energy use, oversee labor practices, and compensate creators.
For investors and entrepreneurs, Hao poses a stark choice:
- Build brittle empires based on exploitation, or
- Construct resilient institutions that promote democratic values and long-term trust.
Why This Story Matters
The rise of OpenAI is more than a business success story. It is a case study in how technological revolutions unfold — and who benefits from them.
In just over a decade, one organization has grown from nonprofit idealism to a $90 billion valuation, reshaping industries and altering global power structures.
Karen Hao’s reporting compels us to confront difficult questions:
- Who owns the future of AI?
- Who pays its hidden costs?
- Do we want to live in a world where unchecked technological empires dominate?
Her response is clear: the future is not inevitable. It is a choice.
