India, Oct. 3 -- Call it artificial or super intelligence, the safety concerns signalled by the signing of the accord by the top AI players - all US-based - OpenAI, Anthropic, Meta, xAI and Nvidia at a dinner hosted by Donald Trump remain largely unaddressed. In this "morally binding" self-policing agreement to prevent AI harm, such harm remains too broadly defined. The jury is still out on whether the harm lies in the technology itself or in the human interface at the labs. Journalist Karen Hao offers an important perspective on this debate in Empire of AI (2025). Hao details the culture that has set the tone for all things AI, including safety and regulation policy: Be the first or perish. Mapping the evolution of AI policy against the rise of the world's premier AI labs - in the book, largely Open AI - Hao shows a shift from openness to secrecy as the stakes got bigger. Empire of AI poses the question: Shouldn't those who stand to be excluded or harmed by AI's march get a say in shaping it?...