- Artificial intelligence is being used in a variety of ways to address some of the Earth’s most pressing environmental challenges.
- Currently, environmental AI involves the analysis of large amounts of unstructured data, but its uses will become more sophisticated as technology evolves and the cost of research decreases.
- It will take a collaborative effort from governments, technologists, investors and business to ensure that AI progresses in an earth-friendly direction.
Emerging technology is being applied in amazing ways to progress humanity. But this doesn’t mean that the Earth should get left behind, even if Elon Musk is bringing us closer to the prospect of moving to Mars.
A series of reports from a collaboration between PwC, Stanford University and the World Economic Forum analysed ways emerging technologies could aid sustainability efforts. The latest piece in the series, Fourth Industrial Revolution for the Earth: Harnessing Artificial Intelligence for the Earth examines the significant opportunity that artificial intelligence (AI) poses for the planet.
The Fourth Industrial Revolution, a concept by WEF founder Klaus Schwab, is underpinned by the digital economy and today’s advances in technology. It also has the opportunity to garner previously inaccessible insights, and enact solutions, to the world’s environmental challenges.
The report suggests that AI – as the technology expected to have the deepest impact – could help transform the sectors and systems needed to address many of earth’s environmental and sustainability issues. As well as ensuring the creation of ‘human-friendly’ AI, something that big tech, government and the corporate sphere have begun to address, society must also ensure that it creates ‘Earth-friendly’ AI, for the sake of our future on this Blue Marble.
Six pressing environmental challenges are illuminated as areas needing priority action: climate change, biodiversity and conservation, healthy oceans, water security, clean air, and, weather and disaster resilience. Here are some of the areas in which AI is being put to use:
Problem: Greenhouse gas levels may be at their highest point in 3 million years.1 Even if current pledges to reduce emissions are met, temperatures are still forecast to be well above the level needed to avoid the worst impacts of climate change.
Potential solutions: Machine learning is being used in renewable energy to enable smart grids, reduce the unpredictability of renewable energy flow and enable better use of resources with lower environmental impacts. Sustainable land-use is being achieved with data-driven farming, crop yield prediction and precision agriculture, improving productivity and limiting wasted resources. Smart traffic lights, parking and on-demand shared transport alongside electric cars and autonomous vehicles are changing the game when it comes to reducing greenhouse gas emissions.
Problem: One in five species on Earth is facing eradication and the planet’s biodiversity is diminishing. By the end of the century, the number could rise to half of all species becoming extinct.2
Potential solutions: AI could help transform the ways in which we monitor and conserve habitats. This may include protecting previous habitats via precision monitoring, improving knowledge of migration patterns, and detection of habitat loss. Drones, in tandem with AI, will enable the identification of poachers, predict routes poachers will take, and allow for animal tracking and the detection of animal capture. The use of AI to identify plant species, register trading of biological and biomimetic assets and monitoring invasive species and diseases will go a long way to understanding what exactly we need to protect and how.
Problem: Oceans are absorbing greenhouse gases,3 leading to acidification and warming with unprecedented flow-on damage to fish and coral.4
Potential solutions: Ocean health is being studied with the use of AI and other emergent technologies. This is allowing issues such as pollution, habitat protection, species protection and sustainable fishing to be addressed. On a practical level, the use of data analytics is providing insight into when patrols should be scheduled in order to catch those fishing illegally. On a larger scale, monitoring marine habitats for changes, such as coral bleaching, or reductions in species quantities will also help understand the bigger picture of ocean life that we know comparatively little about.
Problem: Pollution and climate change are affecting the global water cycle. By 2030, it’s possible that the world would have less than half the amount of fresh water needed to support the global economy.5
Potential solutions: Water security is becoming increasingly important and AI will allow monitoring and management of its supply, efficiency of use both on a residential and urban level, and the detection and monitoring of quality and sanitation. Similar to the concept of peer-to-peer energy lending, there is the potential for the creation of off-grid distributed water resources. By harnessing emerging technologies, local water resources and closed-loop water recycling could become viable. Similar to energy, water could be traded via blockchain with neighbours.
Problem: Air pollution is literally killing people at a rate of one out of every eight deaths.6 Around 92% of the world’s population live in places that fail to meet the World Health Organisation’s air quality guidelines.7
Potential solutions: Machine learning allows for better filtering and purifying, real-time monitoring and forecasting of air pollution. Pollution detection, such as machine learning which analyses photographs on smartphones to detect air pollution levels, will enable individuals and neighbourhoods to adapt their behaviour to current conditions, get real-time alerts and better manage their health. For example, inbuilt sensors are enabling air purifiers to record air quality and environmental data in real-time to adapt their filtration efficiency to individualised needs.
Problem: Climate change has led to an increase in natural loss events, be they geophysical, meteorological, hydrological or climatological.8
Potential solutions: Application of AI to weather and natural disasters is leading to crucial early-warning systems and the real-time communication of danger. With forward planning, urban infrastructure and buildings are more likely to be disaster-ready and the potential damage of a weather event reduced. Mapping, impact and risk analytics are combining to give a greater understanding of how the places we live will be affected in these scenarios, meaning that resilience can be fostered and responses optimised, such as identifying the best routes for evacuation or the amount of relief needed post an event.
As Harnessing AI for the Earth notes, most of the above applications are using AI to unlock value from large, unstructured real-time datasets. As technology advances, data storage and processing power becomes cheaper and artificial intelligence becomes more intelligent, many of the approaches to the above problem areas will increase in sophistication, involving more autonomous, independent AI solutions.
To get to that point, and to ensure that AI progresses towards a future aligned to human and planet-wide values, however, will “require proactive collaboration between policymakers, scientists, civil society, technology champions and investors.”
Ensuring that sustainability principles are part of the conversation of responsible AI will be key. It may not be easy, of course, but, as the report concludes, “if we get it right, we could create a sustainability revolution.”
The 4IR for the Earth programme is a collaboration between the World Economic Forum, PwC, and Stanford University, supported by the MAVA Foundation. The programme looks to accelerate tech innovation for Earth’s most pressing environmental challenges.
John Tomac is a partner at PwC Australia and the leader of PwC’s Sustainability and Climate Change practice.