The book "Empire of AI" (2025) offers a fascinating and in-depth look at the development of OpenAI, from an idealistic nonprofit startup to one of the world's most valuable tech companies. Through journalist Karen Hao's thorough reporting, we learn about the power dynamics, the human costs, and the hidden stories behind ChatGPT's success. This is not just the story of a technological breakthrough, but also a portrait of how ambition, idealism, and commercial interests collide in the battle over the future of artificial intelligence.
Karen Hao is an acclaimed journalist who has worked for the Wall Street Journal and served as senior AI editor at MIT Technology Review. Her articles have been used in academic teaching and cited by governments, underscoring her position as a leading voice on artificial intelligence. With "Empire of AI," Hao demonstrates her ability to combine technical insight with deep social understanding, giving readers a broad perspective on both the advances and dangers related to AI.
Most people have now heard of ChatGPT, and many have experienced how AI has quickly changed everything from healthcare to education, communication, and the workplace. AI has already begun to free us from routine tasks and offers new opportunities to focus on creativity, problem-solving, and human connection. But many also feel unease: What does this development do to us as a society and as individuals? Can artificial intelligence therapy and AI therapists contribute positively, or do we risk losing control over the technology we created?
OpenAI was founded by visionaries like Sam Altman and Elon Musk with the goal of ensuring that AI would benefit all of humanity. At the same time, it soon became clear that the road to that goal was long and winding. The book takes us into the early meetings between Altman, Musk, and other leading technologists in 2015, where idealistic goals quickly had to give way to realpolitik, competition, and the need for capital. When Elon Musk left the project after internal conflicts, it was a sign that the power struggle over AI was already underway. Over time, OpenAI grew into a commercial giant and the focus shifted from community to profit.
Under Sam Altman's leadership, OpenAI was guided by the scaling doctrine—the belief that larger models trained on more data yield better results. Instead of limiting itself to carefully selected datasets, the whole internet was fed into the machines—with all the positive and negative implications that entailed. This required enormous computing power, but also a massive human effort to clean up the most unpleasant content: violence, abuse, hate speech, and extremism. To do this, OpenAI relied on workers in the global South, often paid low wages and working under difficult conditions.
These "data cleaners" were exposed to severe psychological strain, and many experienced trauma as a result of what they had to go through. This aspect of AI development has long been underreported, but Hao devotes significant space to it in the book. She emphasizes that the price of technological progress is often paid by the invisible, and that it is important to acknowledge the human consequences of digitalization. In this way, "Empire of AI" also becomes a story about global inequality and about how power and responsibility are distributed in the digital economy.
To build ever larger and more advanced AI models, OpenAI needed enormous resources. The rescue came in the form of a strategic partnership with Microsoft, which invested one billion dollars in exchange for commercial access to OpenAI's technology. This collaboration changed both the company and the AI industry forever and set off a race to dominate future technology.
When rumors of a strong competitor—Anthropic—began to spread, OpenAI chose to launch ChatGPT at record speed. This was not only a technological breakthrough but also a commercial success: ChatGPT became the fastest-growing consumer application ever. Suddenly everyone—from small businesses to multinational corporations—wanted their own AI solution, and the market exploded. It also put pressure on competitors like Google, Amazon, and Meta to accelerate development of their own AI systems.
Success did not come without costs. Conflicts within OpenAI grew alongside its success. On one side were security researchers who wanted caution and thorough ethical evaluations before launching new products. On the other side were those advocating rapid commercialization and global growth. The split culminated in a dramatic attempt to remove Sam Altman as CEO, where security-focused leaders tried to seize control. After several intense days of unrest, internal revolt, and pressure from both employees and investors, Altman was reinstated and the most critical security leaders left the company.
This episode illustrates how difficult it is to balance responsibility, innovation, and commercial success in an industry where technology evolves faster than society's ability to regulate it. At the same time, it raises questions about who actually holds power over AI—and how that power can and should be governed.
Amid this story of power and technology, it is both necessary and inspiring to reflect on how AI can also be a tool for human growth and support. Especially in mental health, AI's potential has attracted considerable attention in recent years. Digital conversational tools and AI-therapist services have emerged as a supplement to traditional therapy and have already helped many people worldwide.
AI therapy and AI psychologists involve using artificial intelligence to support mental health, whether for prevention, treatment, or follow-up. This can happen in several ways:
These services can make mental health support more accessible, especially for those who hesitate to seek traditional services or live in areas with limited offerings. They can also offer around-the-clock support, which is invaluable for many.
There are, however, significant challenges related to using AI in mental health:
It is therefore crucial that the development, implementation, and use of AI therapists happen with transparency, user focus, and strict ethical standards.
When discussing AI therapy and AI psychologists, the following elements should be considered:
If you are considering using AI therapy, an AI psychologist, or a chatbot psychologist, the following tips may be helpful:
There is little doubt that AI will play an increasingly significant role in mental health. Today, AI therapists and chatbot psychologists are already used in several countries, both in public and private settings. In Norway, several startups and health organizations are testing solutions that combine traditional treatment with digital tools.
We can expect:
At the same time, it is important to remember that the human relationship—and experience—will always be central in mental healthcare. AI should be a tool that strengthens, not weakens, human contact and support.
The story of OpenAI, as told in Karen Hao's "Empire of AI," gives us an honest picture of both the enormous opportunities and the complex challenges that come with artificial intelligence. The book reminds us that technological progress always has a downside, and that power, ethics, and responsibility are as important as innovation.
In mental health, AI therapy, AI psychologists, and chatbot psychologists promise new opportunities for support and treatment. Used correctly, such tools can make help more accessible, personalized, and effective. At the same time, they demand attention to privacy, quality, power distribution, and ethical standards.
By staying informed, critical, and open to dialogue, we can together ensure that AI becomes a positive force in society—a tool to strengthen our mental health, not undermine it. The future is not written, but it is our responsibility to shape it with both wisdom and care.
We hope this article has given you insight into both the big picture behind AI development and the practical aspects of AI therapy and AI psychologists. Wherever you stand in relation to the technology—as a user, professional, or curious reader—the most important thing is to ask questions, seek knowledge, and take responsibility for how we use AI in everyday life.