At a time when the world must cut greenhouse gas emissions precipitously, artificial intelligence (AI) brings large opportunities and large risks. To address its uncertain environmental impact, we propose the ‘Earth alignment’ principle to guide AI development and deployment towards planetary stability.

This book offers a critical primer on how Artificial Intelligence and digitalization are shaping our planet and the risks posed to society and environmental sustainability.

As the pressure of human activities accelerates on Earth, so too does the hope that digital and artificially intelligent technologies will be able to help us deal with dangerous climate and environmental change. Technology giants, international think-tanks and policy-makers are increasingly keen to advance agendas that contribute to “AI for Good” or “AI for the Planet." Dark Machines explores why it is naïve and dangerous to assume converging forces of a growing climate crisis and technological change will act synergistically to the benefit of people and the planet. It explores why AI and associated digital technologies may lead to accelerated discrimination, automated inequality, and augmented diffusion of misinformation, while simultaneously amplifying risks for people and the planet. We face a profound challenge. We can either allow AI accelerate the loss of resilience of people and our planet, or we can decide to act forcefully in ways that redirects its destructive direction.

This urgent book will be of interest to students and researchers with an interest in Artificial Intelligence, digitalization and automation, social and political dimensions of science and technology, and sustainability sciences.

The combination of large volumes of climate and environmental data and analytical methods using machine learning has a long history in research about environmental sustainability. The last years, however, signal a rapid shift in the use of artificial intelligence (AI) for environmental sustainability research and action. AI is however only one part of more profound technological changes. In addition, as the pressure of human activities accelerates on the living planet and the climate system, so too does the hope that a wide variety of technologies that support digitalization such as sensors, advanced satellite technologies, and increased uses of AI will be able to increase societies’ capacities to detect and respond to the repercussions of a rapidly changing planet.

Extreme weather-related events like wildfires have been increasing in frequency and severity due to climate change. Public online conversations that reflect on these events as climate emergencies can create awareness and build support for climate action but are also used to spread misinformation and climate change denial. To what extent automated social media accounts—“social bots”—amplify different perspectives of such events and influence climate change discourses, remains unknown, however. We use Twitter and the 2019/2020 Australia bushfires as a case study to explore this issue. Utilizing more than 1 million Tweets, we identify how climate change is framed in the context of those fires, and to what extent social bots affect specific climate change frames, including the spread of misinformation. Our results show that climate change represents a substantial part of online conversations about fires. The bushfires are primarily framed as a climate change issue including its measurable impacts and political perspectives. Climate denial represents a small share of this conversation and receives limited amplification. Social bots seemingly contribute to the climate change conversation, both through frames that support and oppose climate action, and amplify to larger degree frames appealing to emotions, such as sympathy or humor. We also find that Twitter discussions about the role of social bots in spreading climate denial are amplified more than actual climate-critical frames propagated by bots. A complex interplay between social bots, Twitter conversations, and online news media is emerging, which shapes discussions about climate change and wildfires.

Background: Emerging and re-emerging infectious diseases (EIDs), such as Ebola virus disease and highly pathogenic influenza, are serious threats to human health and wellbeing worldwide. The financial sector has an important, yet often ignored, influence as owners and investors in industries that are associated with anthropogenic land-use changes in ecosystems linked to increased EIDs risks. We aimed to analyse financial influence associated with EIDs risks that are affected by anthropogenic land-use changes. We also aimed to provide empirical assessments of such influence to help guide engagements by governments, private organisations, and non-governmental organisations with the financial sector to advance a planetary health agenda.

Methods: For this integrative analysis, we identified regions in the world where there was evidence of a connection between EIDs and anthropogenic land-use changes between Nov 9, 1999, and Oct 25, 2021, through a targeted literature review of academic literature and grey literature to identify evidence of drivers of anthropogenic land-use change and their association with commodity production in these regions. We only included publications in English that showed a connection between deforestation and the production of one or more commodities. Publications merely describing spatial or temporal land-use change dynamics (eg, a reduction of forest or an increase of palm-oil plantations) were excluded. As we were assessing financial influence on corporate activities through ownership specifically, we focused our analysis on publicly listed companies. Equity data and data about ownership structure were extracted from Orbis, a company information database. We assessed financial influence by identifying financial entities with the largest equity ownership, descriptively mapping transboundary connections between investors and publicly listed companies.Findings 227 public and private companies operating in five economic sectors (ie, production of palm oil, pulp and wood products, cocoa, soybeans, and beef) between Dec 15, 2020, and March 8, 2021, were identified. Of these 227, 99 (44%) were publicly listed companies, with 2310 unique shareholders. These publicly listed companies operated in six geographical regions, resulting in nine case-study regions. 54 (55%) companies with complete geographical information were included in the countries network. Four financial entities (ie, Dimensional, Vanguard, BlackRock, and Norway's sovereign wealth fund) each had ownership in 39 companies or more in three of the case-study regions (ie, north America, east Asia, and Europe). Four large US-based asset managers (ie, Vanguard, BlackRock, T Rowe Price, and State Street) were the largest owners of publicly listed companies in terms of total equity size, with ownership amounts for these four entities ranging from US$8 billion to $21 billion. The specific patterns of cross -national ownership depended on the region of interest; for example, financial influence on EIDs risks that was associated with commodity production in southeast and east Asia came from not only global asset managers but also Malaysian, Chinese, Japanese, and Korean financial entities. India, Brazil, the USA, Mexico, and Argentina were the countries towards which investments were most directed. Interpretation Although commodity supply chains and financial markets are highly globalised, a small number of investors and countries could be viewed as disproportionally influential in sectors that increase EIDs risks. Such financial influence could be used to develop and implement effective policies to reduce ecological degradation and mitigate EIDs risks and their effects on population health.

Online social media such as the microblog Twitter are key digital arenas shaping the public discourse on important societal topics. Automated social media accounts, so-called ‘social bots,’ have emerged as a controversial phenomenon, proven to both disrupt and support online communications on topics such as political elections and public health. To what extent social bots also impact online conversations on climate change, environmental crises, and sustainability remains unknown however.

We present a review of current research on social bots and their potential impact on Twitter discourses around climate, environmental, and sustainability topics; we collect the methods used to detect social bots, approaches to determine their online impact, extract a high-level normative assessment of automation, and summarize the recommendations for stakeholders to manage the challenges created by automation. We note a lack of common, comparable methodologies to robustly assess the impacts of social bots, which contributes to simplistic causal claims about their impact on public opinion and behavior. We identify research needs and offer methodological recommendations for future research on this topic of growing importance.

50 years might not seem long in the history of planet Earth, but the last 50 years have had profound implications for our living planet. 2022 marks the 50th anniversaries of the United Nations Environment Programme and the historic 1972 United Nations Conference on the Human Environment. These anniversaries invite us all to reflect on the successes and failures of collective action and global environmental governance.

Automated decision making and predictive analytics through artificial intelligence, in combination with rapid progress in technologies such as sensor technology and robotics are likely to change the way individuals, communities, governments and private actors perceive and respond to climate and ecological change. Methods based on various forms of artificial intelligence are already today being applied in a number of research fields related to climate change and environmental monitoring. Investments into applications of these technologies in agriculture, forestry and the extraction of marine resources also seem to be increasing rapidly. Despite a growing interest in, and deployment of AI-technologies in domains critical for sustainability, few have explored possible systemic risks in depth. This article offers a global overview of the progress of such technologies in sectors with high impact potential for sustainability like farming, forestry and the extraction of marine resources. We also identify possible systemic risks in these domains including a) algorithmic bias and allocative harms; b) unequal access and benefits; c) cascading failures and external disruptions, and d) trade-offs between efficiency and resilience. We explore these emerging risks, identify critical questions, and discuss the limitations of current governance mechanisms in addressing AI sustainability risks in these sectors.

The COVID-19 pandemic has exposed an interconnected and tightly coupled globalized world in rapid change. This article sets the scientific stage for understanding and responding to such change for global sustainability and resilient societies. We provide a systemic overview of the current situation where people and nature are dynamically intertwined and embedded in the biosphere, placing shocks and extreme events as part of this dynamic; humanity has become the major force in shaping the future of the Earth system as a whole; and the scale and pace of the human dimension have caused climate change, rapid loss of biodiversity, growing inequalities, and loss of resilience to deal with uncertainty and surprise. Taken together, human actions are challenging the biosphere foundation for a prosperous development of civilizations. The Anthropocene reality-of rising system-wide turbulence-calls for transformative change towards sustainable futures. Emerging technologies, social innovations, broader shifts in cultural repertoires, as well as a diverse portfolio of active stewardship of human actions in support of a resilient biosphere are highlighted as essential parts of such transformations.

Globally, financial services are well positioned to contribute to the transformation needed for sustainable futures and will be critical for supporting corporate activities that regenerate and promote biosphere resilience as a key strategy to confront the new risk landscape of the Anthropocene. While current financial risk frameworks focus primarily on financial materiality and risks to the financial sector, failure to account for investment externalities will aggravate climate and other environmental change and set current sustainable finance initiatives off course. This article unpacks the cognitive disconnect in financial risk frameworks between environmental and financial risk. Through analysis of environmental, social, and governance ratings and estimates of global green investments, we exemplify how the cognitive disconnect around risk plays out in practice. We discuss what this means for the ability of society at large, and finance in particular, to deliver on sustainability ambitions and global goals.