Will artificial intelligence (AI) be the panacea of environmental governance in an age of planetary crisis? Can technology save us from ourselves? We already seem unable to curb our enthusiasm for consumption, which is further encouraged by “AI fueled advertisement … set to turbocharge consumerism” (180). This is a core exploration in Peter Dauvergne’s book AI in the Wild. He examines the role of AI as a contributor of data, precision, trust, and efficiency of governance, in language that is accessible and at times reads like a novel, making you want to know more and keep turning pages. The examination spans cross-sectoral topics from land to sea and considers the pros, cons, and in-betweens of the birth and rapid growth of AI in businesses, education, homes, environmental governance, and everywhere else.

And you come to realize this almost immediately—AI is in everything we do, everywhere we go, all at once. And it can be used for amazing things, as Dauvergne points out, highlighting some of the genius ways in which AI is used precisely to advance humanity and rectify some of the damage we have inflicted on our environment. But the benefits of AI can also be used for all that’s bad in the world, and—spoiler alert—an important conclusion of the book is that AI will not save us. For Dauvergne, there seemingly is no way out of our current state of affairs within the contemporary global order, because AI cannot “overthrow the entrenched interests that are exploiting people and nature” (8). Though this impression permeates the book, Dauvergne does try to balance the book on the good side of the edge of despair as he also guides the reader through the benefits—and exploitation potentials—of AI.

The book follows the use of AI in several different sectors spanning the themes of conservation, ecobusinesses, smart products, and smart cities and farms. It follows a similar pattern in each example—starting with emphasizing how it can benefit governance of a given sector with its use. AI can monitor illegal logging of rainforests to alert managers faster, eradicate invasive species harmful to a given ecosystem using underwater robots with machine vision technology, empower police officers to identify and arrest ivory poachers in Africa with camera catch technology in parks, reduce greenhouse gas emissions with self-driving, AI-powered electric vehicles—the list goes on, and the scale tips the reader toward careful optimism for the future. Dauvergne then goes on to caution against this optimism, identifying the possible limits and dangers of AI in ways that benefit those who are more privileged, arguing that all of this mostly leads to everyone everywhere buying more of everything instead—and green growth taking place in addition to the standard growth, not as a replacement for it.

Dauvergne also emphasizes that AI is allowing us to gain access to more and more data and knowledge, though quoting an AI scientist who emphasizes that in deep learning, “there’s no data like more data” (54). For the ocean, Dauvergne discusses AI tracking of fishing vessels, for example, giving us access to comprehensive data about their movement. This argument also points to a weakness in his analysis, however. Throughout the book’s examples, it feels like we are only skimming a superficial surface and not diving into some of the real advances made. The examples do not acknowledge or explore existing regulation, perhaps a testament to the fast pace of both AI and mechanisms that have come to regulate it in recent years—including out of ethical concern. AI can make profound changes to how governments work and cooperate to bring the state of the environment back within planetary boundaries, as Dauvergne acknowledges. In fisheries, for example, AI-driven automatic catch registration regulations coupled with port state measures can eliminate illegal fishing activities by preventing illegal fishers from landing their catches.

These shortcomings aside, AI in the Wild is a thought-provoking and—unusual for books in political economy—thrilling read. It not only contributes to the global environmental politics and international political economy literature but also is a source of accessible literature about AI more generally, removing the term from the ownership of science fiction and tech books. After all, we are surrounded and affected by AI in our daily lives—with smart home appliances like Alexa or Siri, automatic lightbulbs, and facial recognition on our phones. Future research can use this book as a springboard to dive deeper into AI within specific sectors and explore the value chain throughout it. Others can further explore the use of blockchain technology as a nongovernance ally of environmental governance and assess its use as a transparency tool for ensuring that AI is a tool not only of the supervillains of the world but also of the masses, who would then have access to this information in a chain of data that all can trust.

Author notes


SINTEF Ocean, Trondheim, Norway. Contact: Rachel.Tiller@sintef.no