Is AI Good for the Planet?. Benedetta Brevini
Читать онлайн книгу.
is also a significant amount of research connecting AI systems to applications in medicine. An international study examining the potential use of AI in the field of dermatology has found improved diagnostic accuracy and clinical decision-making when AI is used in conjunction with human clinical checks, suggesting stronger results than the use of AI or experts alone (Tschandl et al. 2020). The artificial neural networks analyse uploaded pictures and identify potential sites of malignant melanoma; this is currently being tested in an Australian skin clinic. The program has since been able to recognize melanoma at sizes smaller than the human eye can detect, for instance melanomas as small as 0.2 millimetres (UQ News 2020).
This is without mentioning AI’s current applications in the global conservation field – from modelling biodiversity loss to monitoring migratory species and climate change scenarios. AI’s emerging use in agriculture has aimed to optimize everything from crop production to resource consumption, harvesting, monitoring and processing. In 2019, twenty-five European countries signed a declaration of cooperation for the digitalization of agriculture, in an acknowledgement of the potential power of digital technologies and with the goal to establish infrastructure designed to support a smart agri-food sector (European Commission 2020a). Projects funded by the European Union (see Brevini and Murdock 2017) include Sweeper, a sweet pepper-harvesting robot that uses algorithms for fruit detection and localization and that was the first of its kind to demonstrate harvest success in a commercial greenhouse (Arad et al. 2020).
Or take CORaiL, an AI-powered solution to monitor and analyse coral reef resilience. Since May 2019, it has been deployed in the reef surrounding Pangatalan Island in the Philippines and has helped researchers to study the effects of climate change in the area (Wu 2020). Another famous robot wandering around the Great Barrier Reef in Australia is called LarvalBot and designed to carry coral larvae across destroyed areas of the reef. The larvae are distributed so that new coral colonies can form and new coral communities can develop. This process mitigates the damage caused by mass bleaching, weather events and climate change (Cimons 2019).
Some of these AI technologies are simply processing the data according to a basic set of formulae. Others are more complex systems that are effectively able to teach themselves progressively and learn from data as they are collected.
Despite the different complexities and endless applications, the dominant rhetoric around AI extends far past its current capabilities. Accounts across countries throughout the world proclaim the imminent development of intelligent machines, capable of outsmarting the human mind, amid promises to change everything fundamentally, from our working lives and domestic habits to transport and health services – to name just a few areas that will be affected. In the last decade we have witnessed a clear increase in predictions that the arrival of superintelligence is imminent; thus nations are expressing an urgent need to be ready for AI. As Goode writes, this is leading to the ‘sublime spectacle of inevitability … that does little to offer lay citizens the sense that they can be actively involved in shaping its future’ (Goode 2018: p. 204).
AI is thus being promoted as the principal solution for many of humanity’s challenges; it is put forward as inevitable and ineluctable (Brevini in press). The following statement, taken from the official communications of the European Union, could not be clearer: ‘AI is helping us to solve some of the world’s biggest challenges: from treating chronic diseases or reducing fatality rates in traffic accidents to fighting climate change or anticipating cybersecurity threats’ (European Commission 2018b).
The High-Level Expert Group on Artificial Intelligence (AI HLEG) appointed by the European Commission goes into even greater detail about the capabilities of AI to make humanity ‘flourish’ by solving virtually all problems of society:
We believe that AI has the potential to significantly transform society. AI is not an end in itself, but rather a promising means to increase human flourishing, thereby enhancing individual and societal well-being and the common good, as well as bringing progress and innovation. In particular, AI systems can help to facilitate the achievement of the UN’s Sustainable Development Goals, such as promoting gender balance and tackling climate change, rationalising our use of natural resources, enhancing our health, mobility and production processes, and supporting how we monitor progress against sustainability and social cohesion indicators. (European Commission 2019a)
Consequently, when the artificial machine arrives in this future–present that is always inevitably imminent, it will manifest as a superior intelligence, capable of solving the problems that humans have created – and climate change in primis.
As citizens, we are almost left with the sense that this artificial entity that will come to rescue humanity, the world, and all living things is a divine, magic hand, a deus ex machina.
This portrayal of AI as a benevolent deity has a crucial effect: it obfuscates the materiality of the infrastructures and devices that are central to AI’s functioning. In all its variety of forms, AI relies on large swathes of land and sea, vast arrays of technology, and greenhouse gas-emitting machines and infrastructures that deplete scarce resources through their production, consumption and disposal. AI requires increasing amounts of energy, water and finite resources.
Why are we not talking about the negative impact of AI on the climate crisis? This is precisely what I want to discuss in this book. And more: I want to bring the climate crisis to the centre of debates around AI developments.
Clearly there are other important concerns about AI developments, from moral and ethical appeals for caution in the use of AI in military operations to mounting fears in areas where human expertise is crucial to safeguarding human rights (such as public health and the judiciary). There are huge ethical issues concerning documented algorithmic racial and gender biases, and fears that AI will make human labour redundant, producing a class of supereducated employees and another of less educated, unemployable workers.
These concerns are beyond the scope of the present book. If we lose our environment, we lose our planet and our lives. So we must understand and debate the environmental costs of AI.
The COVID-19 global pandemic caused the worst economic contraction since the Great Depression. It underscored the need to rethink the type of economy and society we want to build as we face the worsening climate crisis. Bold green recovery plans have been announced all around the globe (European Commission 2020b), and we have been told that technological innovations and AI are at the centre of recovery, as they have the potential to create millions of jobs and to boost economies devastated by the pandemic. But what would happen if we discovered that the environmental impact of AI is so massive that it compromises the plans to decarbonize the economy proposed by Green New Deals around the world?
The climate crisis is here to stay
In October 2018, the world’s leading climate scientists who comprise the Intergovernmental Panel on Climate Change (IPCC) warned the world that there were only twelve years for global warming to be kept to a maximum of 1.5-degree C (see IPPC 2018). Beyond that level, the risks of life-threatening weather events such as drought, wild fires, hurricanes, extreme heat and poverty for millions of people will be significantly worse. One need only look at the devastation of Australia’s ecosystem by the bushfires in the summer of 2019 to understand the gravity of the situation.
In November 2019 we received even clearer warnings from the UN Emissions Gap Report 2019, which assessed the gap between anticipated emissions in 2030 and levels consistent with the 1.5°C increase in temperature outlined in the Paris Agreement (UN Environment Programme 2019). It is estimated by the IPCC that an increase in average global temperature beyond this limit will lead to loss of habitable land for many humans and other life forms, as well as to catastrophic water shortages and crop shortages; this will cause massive migration and possible conflicts, as populations seek safer ground. The UN report also explains that the national commitments made in Paris must increase at least fivefold if we are to prevent a temperature increase greater than 1.5°C. Unless emissions fall by 7.6 per cent each year in the period between 2020 and 2030, the world will miss the opportunity to limit the damage.