Eleanor Drage's AI Ethics Book Reviewed

AI ethicist Eleanor Drage, in her book "What If We Got AI Right?", posits that for humanity to coexist effectively with artificial intelligence, a crucial step is recognizing AI's inherent "humanity". Drage clarifies this by explaining that AI is not an abstract, ethereal entity but rather the direct outcome of extensive human effort. This labor spans from the extraction of materials in silicon and quartz mines to the intricate processes within microchip factories and the meticulous work of data labeling centers. She contends that citizens seeking to reclaim power from the dominant tech companies developing AI must possess a clear understanding of what these companies are marketing. This understanding necessitates piercing through the pervasive marketing jargon and hype. For instance, the concept of "the cloud," often perceived as a distant, nebulous storage space for digital assets like photos, is, in reality, simply another entity's computer infrastructure. Similarly, AI "hallucinations" are reframed not as mysterious AI behaviors but as system errors or mistakes in data labeling. Drage also challenges the broad and potentially misleading application of the term "intelligence" to AI systems.

Drage critically examines the prevailing narrative in Silicon Valley, which often presents AI as a binary force: either a harbinger of unprecedented prosperity and extended human lifespans, or an existential threat leading to global annihilation. She argues persuasively that the major technology corporations are fundamentally incapable of fulfilling their utopian visions. This inability stems from their persistent prioritization of profit and power over societal well-being, and their tendency to serve the interests of a select few. Concurrently, Drage highlights how the cultural fixation on an AI-induced apocalypse serves as a significant distraction. This focus diverts attention from more immediate and practical discussions concerning AI safety. Such conversations could involve empowering citizens with greater control over their personal data usage and influencing the training and regulation of AI models. Furthermore, Drage suggests that the apocalyptic scenarios propagated by tech leaders create a false sense of urgency, elevating AI safety above other pressing global issues. These issues, including climate change, geopolitical instability, and economic inequality, are already inflicting substantial suffering and, if effectively addressed, could inherently mitigate many AI-related risks.

Drage's work encourages a shift in perspective, moving away from abstract fears of AI dominance towards a more grounded understanding of its creation and implications. By demystifying AI and highlighting the human labor involved, she aims to empower individuals and foster a more informed public discourse. The book advocates for a proactive approach to AI governance, emphasizing the need for transparency and democratic oversight in its development and deployment. This contrasts with the often top-down approach favored by large technology firms. Drage's critique of the profit-driven motives within the tech industry underscores the potential for AI to exacerbate existing societal inequalities if not carefully managed. Her call for recognizing AI's "humanity" is a call to acknowledge the human systems and decisions that shape its capabilities and impact, thereby enabling more effective and equitable integration of AI into society.