[Lee Myung-ho's Future Preview] Can AI Help Solve Sustainability Issues?

COVID-19 tragically reminds us that nature can be ruthless. Over 100 million people have been infected and more than 2 million have died within a year. Vaccines, which were expected to take several years to develop, were created in less than a year and vaccination began, leading to a decline in the spread of COVID-19. However, developing countries still lack access to vaccines, so humanity’s suffering seems likely to continue for a longer time. The phrase “a life without good politics, trustworthy science, and strong technology is lonely, poor, dirty, cruel, and short” comes to mind. It reminds us that only when good politics lead the people in the right direction, trustworthy scientific knowledge prevents confusion and division, and strong technology finds solutions to problems can human life improve.

We will overcome COVID-19, but can we also overcome the gradually approaching climate crisis? Until recently, some scientists claimed that we are currently in a mini ice age and that global warming was a hoax. This claim is no longer accepted, and the scientific community now acknowledges global warming as a fact. Scientists estimate that if global warming continues, agricultural production worldwide could decrease by up to 30% by 2050, with 500 million farmers in developing countries being the most affected. The number of people lacking sufficient water for one month each year could reach 5 billion, and hundreds of millions of people could be displaced from coastal cities due to rising sea levels and storm surges. Many people might doubt these simulations and forecasts. They want to avoid these issues, thinking they involve knowledge they do not have and an invisible distant future. Trust in science remains fragile.

Relying on this avoidance mentality, bad politics still expresses doubts about global warming or shows a lukewarm attitude. There are political declarations, but no willingness to bear the costs and pain that come with them. Achieving carbon neutrality within 30 years is estimated to require an investment of about $13.7 trillion, equivalent to approximately 16% of the world’s GDP. If strong carbon regulations are implemented, typical consumer goods companies could lose up to 20-25% of their current margins, and if the carbon price per ton surges from $20 to $75 by 2030 (the level proposed by the IMF), corporate management will become difficult. No government finds it easy to pursue policies that hinder economic growth and impose high costs on citizens. Of course, there are arguments that strong investments and regulations will bring benefits that far exceed the losses caused by the climate crisis in the future, but the current pain is significant and the “hell” of the future is not visible. Can good politics gain the support of the people?

COVID-19 has shown how great the economic pain can be to reduce carbon emissions. Lockdowns and social distancing reduced economic activity, making the air cleaner, but extreme weather events worldwide simultaneously showed that global warming cannot be solved with that level of pain. However, the biggest problem is the lack of strong technology to solve it. The 2015 Paris Climate Agreement declared that carbon neutrality must be achieved to prevent further warming of the Earth. Achieving a state where carbon emissions and removals offset each other to reach zero emissions requires strong technology. Comprehensive efforts are needed, from renewable energy and energy-saving technologies to lifestyles that conserve energy.

Meanwhile, the experience of COVID-19 has raised expectations that digital transformation could contribute to carbon neutrality. In particular, expectations for artificial intelligence (AI) are growing. Can AI technology become a powerful technology to solve sustainability issues? There are still many hurdles to overcome. Training the most advanced writing AI, GPT-3, emitted 300 times the carbon of a person flying from New York to San Francisco, and five times the lifetime carbon emissions of a car in the United States. As digital hardware and software, including AI, increase, so does the carbon footprint. AI is especially energy-intensive because it processes large amounts of data computationally. This is why technologies that reduce computing power, such as edge AI chips and neural network compression, are gaining attention.

AI also demonstrates excellent effects in energy saving. DeepMind, developed by Google, which created AlphaGo, reduced energy use in data centers by 35%. In addition, attempts to use AI for energy saving are increasing in many fields. These include reducing waste and energy use in manufacturing processes, maximizing heat recycling and heating/cooling efficiency in buildings, improving traffic congestion and freight transport networks, coordinating renewable energy supply and demand through weather forecasting, increasing crop yields while reducing fertilizer and water use, and discovering new materials. Heavy chemical industries account for 30% of the world’s total annual carbon dioxide emissions and are expected to increase to 46% by 2050, so reducing carbon emissions in this sector is crucial.

Overall, AI can contribute to sustainability in three ways. First, by reducing errors, it saves the energy and costs required to correct mistakes. Second, it improves efficiency by helping redesign optimal processes. Third, it enables integrated resource management, managing the life cycle from raw materials to products to prevent resource loss and environmental pollution. The circular economy model has the potential to reduce global emissions of aluminum, plastic, steel, and cement by 40%, and by 56% in developed countries by 2050.

When the internet was introduced in 2000, companies thought hiring a webmaster was enough. However, disruptive innovation was the emergence of digital business models. For AI to solve sustainability issues, it must be approached through the reorganization of business processes and models.

Myungho Lee, Planning Committee Member, YeoSiJae Foundation