AI can drive global GDP and cut GHG, says PwC and Microsoft
PwC and Microsoft have released a report suggesting AI-driven technologies can create both GDP growth and GHG reduction
Artificial Intelligence is being promoted as the next great driver of economic growth, but new research conducted by PwC on behalf of Microsoft suggests that not only could AI create economic value, but it could also be applied in parallel to create considerable reductions in global carbon emissions.
The research, published in a whitepaper, How AI can enable a Sustainable Future, suggests that AI-enabled solutions and improvements in four key areas — agriculture, water, energy and transport — could see both a boost of 3.1% to 4.4% in global GDP, representing up to $5.2 trillion in value, coupled with a reduction in greenhouse gas (GHG) emissions of 1.5% to 4%, as much as 2.4 gigatons
The report notes that PwC’s own estimates already suggest that AI could contribute as much as $15.7 trillion to the global economy by 2030, but that previous industrial revolutions and technology advancements have often come at the long-term cost of degradation to climate and environment. One assessment says that the cost of climate change to the US could reach over $500bn a year by 2090 if no efforts are taken to mediate climate change.
The research is intended as a preliminary assessment of the potential of AI for impact on economic growth and emissions reduction potential, between now and 2030.
Writing in the introduction to the report, Celine Herweijer, Global Innovation & Sustainability Partner, PwC UK and Lucas Joppa, Chief Environmental Officer, Microsoft, said: “Microsoft and PwC have a shared belief that, with a great deal of urgency, we must address the diminishing health of our environment. We also share a belief that new technologies like AI can be a game-changer in this space. But to create the kind of global action needed, it needs to be more than a belief — it needs to be backed by data, by solutions, and by new partnerships.”
The report authors says they have focused on the four sectors which are equally important to the economy and to the environment. They note that the gains given in the report are against a ‘business as usual’ scenario, but that there are many initiatives to reduce carbon emissions, independent of technology, which are already underway.
AI CUTTING CARBON, DRIVING GROWTH
The AI-driven gains are mainly achieved through greater productivity output, and automation of tasks, and rely on a combination of artificial intelligence and complementary technology and infrastructure, such as robotics, IoT, distributed energy resources, electric vehicles and so on.
In terms of environmental benefits, the report considers not just the greenhouse gas reductions, but also potential improvements in areas like air and water quality, prevention of deforestation and protection of biodiversity.
Among the potential scenarios that combine economic growth with environmental impact in the energy sector are deployment of AI-enabled distributed energy grids, that are able to make maximum efficiency from distributed energy generation and storage, smart meters and dynamic pricing.
In transport, AI-enabled autonomous vehicles can be far more efficient in terms of vehicle performance through better driving behaviour and/or less miles driven by more efficient route planning, but can also enable carbon reductions through enabling ride-sharing schemes.
In the agriculture sector, the report suggests that AI can be used to analyse satellite data and ground-based sensors to monitor forest conditions in real-time and at scale, providing early warning systems for investigation of illegal deforestation, with the potential to save 32 million hectares of forest globally by 2030.
Air pollution is one of the largest environmental risks to human health, where using AI to provide more accurate and localised early-warnings of poor air quality can help reduce this burden. PwC estimates that using AI in this way could provide additional economic benefits of $150m globally in 2030 in reduced healthcare costs and health impacts.
However, the report authors also note that AI-driven solutions will rely on intensive computing, so failure to use clean energy for compute resources, especially in data centres, will mean that AI-driven gains will potentially be wasted.
“There is enormous potential for AI to be an important tool in the effort to decouple economic growth from rising carbon emissions,” Herweijer and Joppa wrote.
“This outcome relies on bringing together several factors. The solutions we explore are not AI acting on its own; in most cases multiple complementary technologies come together. AI also requires large amounts of compute power, which translate to energy consumption. Without new incentives that accelerate a market change towards clean energy — from renewables to electric vehicles — the efficiency gains from AI won’t deliver their full emissions reduction potential for the world.”
The headline finding of the report is the potential of using AI in environmental applications across the four verticals could boost GDP by as much as 4.4% or $5.2 trillion, and reduce greenhouse gas emissions by as much as 4%, equivalent to around 2.4 gigatons of CO2 — the equivalent of the predicted 2030 carbon emissions of Australia, Japan and Canada combined.
PwC also predicts that the AI applications in its analysis could create between 18.4 and 38.2 million jobs worldwide, particularly in skilled roles, up to a 1% net of total global employment in 2030.
The sectors with the greatest potential for GHG emission reductions at energy, which could reduce emissions by up to 2.2% and transport, up to 1.7%, although the report notes that enhancements in agriculture and water will have positive benefits in other areas of the ecosystem and environment, and are also particularly important in terms of food and water security and quality of water supply.
One factor that PwC and Microsoft stress is that the economic benefits of AI-environment applications will not be felt equally around the globe. It is primarily the regions that are already ahead on technology deployment that will gain the most. Europe, East Asia and North America could each make GDP gains in excess of $1 trillion — the best GDP gains would be in Europe, which could get a total GDP increase of 5.4%; while North America could cut GHG emissions by 6.1%.
In the MENA region, GDP gains could be as much as 2.9%, and a 1.7% carbon emission reduction.
The worst performing regions will be Latin America with a potential 2.2% GDP gain and 0.9% GHG reduction, and Sub-Saharan Africa, with only a 1.3% GDP gain and 0.2% GHG reduction. The report notes that these lower performing regions, which are less affluent, are also most likely to suffer the most negative physical impacts of carbon gains.
The picture will be the same in the distribution of new jobs created by environmental AI applications, which will be concentrated in the developed regions, which could lead to further inequality. Imbalance in job opportunities could also spur more migration.
Herweijer and Joppa said: “There are important issues of justice to consider, to ensure that benefits are inclusive. The largest gains we see map to the countries that are already at the forefront of AI adoption, and are not evenly experienced today. Without incentives and policy change to ensure all regions are ready to capture these benefits, these economic and climate inequalities will be exacerbated.
“Just looking at jobs, the good news story of more high-skilled jobs also carries a reality of jobs displacement, and a pressing need for upskilling and reskilling to avoid leaving people behind.
“This all means we need to think beyond the technology itself to address the wider implications on society and our environment. From the need for strong ethical frameworks, to the evolution of laws, the importance of education and training for new skills, and even labour market reforms — these must all come together if we’re going to make the most of this new technology.”