Empire of AI: Inside the Reckless Race for Total Domination by Karen Hao documents the story of OpenAI and the people who built it – a story of imperial ambition and aggrandisement whose effects have exploded into society (whether we like it or not). Hao tracks OpenAI’s path and the story of AI research and development from open scientific research to being a for-profit tech company pursuing growth at all costs and fuelling what could very well be an economic bubble.
She also tracks the career of OpenAI’s most famous individual, Sam Altman, and his path of being one among many to OpenAI’s head and pushing out many of the key people who helped to build OpenAI – centring the company around himself and loyalists who would follow his approach of building the company for profit and pursuing growth at all costs. Hao shows how Altman’s public persona has not matched how he has worked within OpenAI and his willingness to push past safety concerns and the impacts the company’s work could and is having on society is worrying (to say the least). Hao’s access from OpenAI’s early days means Empire of AI has some great insights into how the company came to be, how it grew, and most importantly, how it changed as it grew and started releasing ChatGPT out into the world.
She also highlights the impacts that the AI boom is having around the world – its reliance on poorly paid workers in the Global South to annotate data, the impact of ever expanding data centres in size and number, the sheer volume of data, power, and water that generative AI requires and how this is used, and the impact of increasing demand for computer components. These impacts are often not acknowledged or are downplayed by generative AI companies and AI advocates or justified as necessary evils in the pursuit of the ‘noble goal’ of artificial general intelligence.
Hao’s skepticism of OpenAI and her analogy of OpenAI and its counterparts as imperial beings, exploiting the resources of the world to further enrich themselves, and exporting the consequences of their work to others was an interesting and persuasive way to frame her book. She also explores ways in which people and communities are challenging this and different models and examples of how things could be. She also has a great ability to explain how generative AI models are built and operate to the layperson, especially helping to sift through the public relations and (let’s be honest) propaganda of generative AI companies and their fellow travellers to understand what these models are and what they can do.
Hao’s book is an important read and learning opportunity – to reflect on how this rapid growth and change in society is working, and whether it is something we want at all, because despite OpenAI’s claims and wishes, it is not inevitable. A useful opportunity before we have to have our very own Butlerian Jihad.
Karen Hao will be speaking at an upcoming WORD Christchurch event on Monday 11 May.
Add a comment to: Empire of AI: Inside the Reckless Race for Total Domination