Penguin Press
Empire of AI: Dreams and Nightmares in Sam Altman’s OpenAI by Karen Hao
Published in May 2025
The AI book that I’m currently recommending to my higher ed colleagues is Karen Hao’s Empire of AI.
Nowadays, all we can talk about at every gathering, on and off campus, is AI. Together, we worry about the role of AI in academic integrity, student learning and authentic assessment. We experiment with how AI might enhance productivity in course design, program marketing, technology services and administrative processes. We talk about how AI tools are changing how we approach our daily work as faculty and staff, with varying levels of enthusiasm and trepidation.
When we talk about AI on campus, we mostly mean generative AI. But what if we are having the wrong—or, if not wrong, then incomplete—campus conversations about AI?
What makes Empire of AI an exceptional book is that Hao elevates the story beyond a single corporate biography of the company OpenAI to a broader consideration of AI’s impact on global workers, the environment and the future of the labor market. Generative AI is not a clean technology from an environmental or a human/worker rights standpoint.
But what if generative AI is only a brief stepping-stone to artificial general intelligence? This is a trajectory that the researchers, technologists and entrepreneurs profiled in Empire of AI share a near-universal belief in. For the people working on frontier AI models at companies like OpenAI, Anthropic and Google, AGI is years—not decades—away.
It is easy to forget that OpenAI was originally created as a nonprofit to research and develop mechanisms to mitigate the risks of AGI. The book traces OpenAI’s transition under its CEO, Sam Altman, from a mission-driven research lab to an enormous for-profit company, with a market cap of over $850 billion at the time of this writing.
Hao is very good at integrating the corporate intrigue of AI, including the brief firing and rehiring of Altman (the Blip) and the creation of Anthropic (the Divorce) by past OpenAI employees, including the siblings Dario and Daniela Amodei.
What I had not fully understood until Empire of AI is the extent to which the people at the forefront of AI research think about the timeline of AGI. Artificial general intelligence, a state of development at which the technology reaches human-level abilities across all the domains humans operate in, has always seemed more science fiction than reality.
The Star Trek: The Next Generation character of Data is fun to think about, but the potential that Data would be competing for my job (and every other job) in my lifetime seemed remote. Now, after reading Empire of AI, I’m not so sure.
What would it mean if those of us focused on the university as a workplace took seriously the potential that something approaching AGI might arrive within the next decade?
How would AGI change the composition of our work and workforce?
Are universities immune from the dramatic shifts in demand for knowledge workers that we are starting to see early indicators of in other industries?
If AGI is set to arrive within the decade, as those in the AI industry think will occur, then what should we be doing at our universities to prepare for this change?
What are you reading?
