Dichotomy with Artificial Intelligence for Climate Change | Data Driven Investor | by Deepa Ramachandra | April 2024

Image source: iStock

“We need an AI race: to develop artificial intelligence that… strengthens climate action (and) artificial intelligence that moves us towards the SDGs.” — UN Secretary General António Guterres

Today’s climate crisis thus says that a rush to find any solution that “overburdens” climate action is inevitable. The latest artificial intelligence hysteria has made it a major technology for solving all the world’s problems, including climate change. Although artificial intelligence may transmogrify as an effective solution to all climate problems, we should not discount its environmental costs. To resolve this dichotomy, a balance needs to be struck, weighing the help and harm that climate AI can cause.

The solution to the climate problem lies in understanding the historical trends of change, analyzing its causes and obtaining practical knowledge. It would mean collecting vast amounts of complex data – from cameras to sensors to satellites – and making sense of it. This is where AI shines.

Predict future climate patterns

AI models are widely used today predict future climate patterns. Regional modeling of sea level rise, wildfires, floods and droughts helps predict extreme weather events by location. Insights from AI systems would provide data-driven information on rising sea levels and impending weather hazards to policymakers and help develop accurate early warning systems to alert communities. They can better plan infrastructure investments to build disaster-resilient buildings.

Forecasting the supply of renewable resources

Machine learning models are also accurate in advance of how the supply of renewable resources will be, such as wind and solar power, will be another day. By combining historical energy production data with current weather information and synthesizing missing data, these models provide timely and reliable forecasts of renewable energy. Accurate forecasts of wind or solar power that would be available when given hours in advance, thanks to AI, would help authorities plan for energy supplies. They are thus ready to cover all energy demands on a given day without downtime. Optimized grids go a long way in achieving that Sustainable Development Goal (SDG) 7 — Ensure access to affordable, reliable, sustainable and modern energy for all.

Optimize carbon offsets

Artificial intelligence also plays a role in optimizing carbon offsets. From the vast amount of data, it can detect the impact of each carbon emitter and remotely measure natural carbon stocks. AI is also effective at the micro level, as it can calculate the carbon footprint of individual products. The models could assess carbon capture storage and improve carbon sequestration practices. It can alert people to dangerous levels of air pollution in a particular area monitoring urban air quality reports. The use of artificial intelligence in carbon markets would help set carbon prices accurately and backed by data. Additionally, monitoring air quality using artificial intelligence would provide information to improve urban planning, public health, transportation and waste management.

Improve crop yields

In addition, AI models could predict the best planting times by combining weather insights, renewable energy forecasts and climate influences. And thereby improve crop yields. They could assess soil health and monitor pest and disease outbreaks. Smart irrigation systems powered by artificial intelligence could reduce water consumption at a time when droughts and famines are becoming more frequent and protracted. With improved modeling of climate change patterns, communities can plan effective strategies for better crop yields.

Optimize supply chains

Another area where artificial intelligence benefits the climate is the reduction of industrial waste optimization of supply chainsmonitoring resource consumption and planning recycling efforts.

Prevent biodiversity loss

Finally, the emerging AI market can help prevent biodiversity loss. Machine learning models help monitor encroachments on forests and nature reserves. They are used to identify and regularly count different species. As a result, we are able to predict their large-scale migration patterns. Monitoring biodiverse life forms with AI it predicts the likelihood of species extinction and therefore gives policymakers an edge in creating effective strategies to conserve and restore the planet’s biodiversity.

As with any technology, AI is not a panacea for all climate problems. In reality, this touted technology comes with its own burden of climate problems.

Image source: Tasks and average CO2 emissions for 1,000 queries

The carbon footprint of artificial intelligence models is deadly enough that discussions about using them to tackle climate change are often deadlocked. Although environmental damage depends on several factors, including model type, size, number of parameters used, amount of data processed, and data centers running the models, it is not uncommon to categorically associate AI models with harmful carbon emissions. The result is a desire to frequently release faster and more complex hardware to improve model accuracy intensive cycling through hardware, renders a lot of hardware very quickly very useless. But with hardware recycling still struggling to catch up, this e-waste fueled by the booming AI industry is taking a toll on the environment like never before.

Diving further to make AI the answer to climate disruption is a condition of existing solutions. Current AI-based climate change remedies are sparse and difficult to access and scale. Except, AI is being used in the fossil fuel industry to accelerate oil and gas production and inflate their coffers. As a result, AI often gets bogus answers from climate advocates.

The use of artificial intelligence to solve the climate crisis while minimizing its adverse effects on the environment can be done in several ways. It might start with considering whether artificial intelligence is a must for a climate project – if it’s unnecessary, leave it out. Instead of developing a new model each time, it makes sense to use existing models and adjust them according to the underlying data and conditions. Smaller models are more energy efficient than their larger counterparts. Encouraging the use of renewable sources to power these models would further reduce environmental damage. Improving stakeholder engagement would contribute to better recycling practices. Access to capital and trained professionals could bridge the gap between academic research and widespread adoption of universal and inclusive AI models. Instead of a fragmented approach integrating artificial intelligence for climate into national policies would help plan a comprehensive global AI strategy for climate action. Scientific bodies under the auspices of governments should come up with standards for AI safety reporting. Individuals and companies should be held accountable for non-compliance.

In conclusion, despite the mixed impact of machine learning on climate change, there are some benefits to using AI to address the climate crisis. Artificial intelligence, if used correctly, can be a catalyst for achievement SDG target 13 — Climate action and the fight against climate change and its impacts.

Leave a Comment