Artificial Intelligence (AI) and Machine-Learning Programmes have emerged as the front-running technologies in the bid to tackle carbon emissions and climate change. New initiatives aimed to improve efficiency and reduce waste are being rapidly developed and are seeing promising results. However, the carbon emissions of these products cannot be ignored. In this blog we will explore if machine learning can help us to be greener, or if their own carbon-emissions price is too high to pay.
Machine Learning: A Definition
According to computer scientists, machine learning is “a form of artificial intelligence that enables computers to use historical data and statistical methods to make predictions and decisions without having to be programmed to do so”.
In day to day life we all encounter machine learning, perhaps without realising it. AI applications, such as Google Translate, email spam filters, and predictive text functions all use it, and more and more technologies using machine learning are being developed as we speak. This includes machine learning technologies in the field of climate research.
Machine Learning and the Climate
Data scientists and climate researchers in Canada are currently working together to use machine learning to monitor and track changes to the climate, and the public’s response to it.
Some of the technologies that have recently been developed include technologies that track the development of climate change in the country, with the ability to warn farm owners about potential crop damage as our weather systems become more unpredictable. Climate forecasting is arguably emerging as one of the major technologies in the country.
On a domestic scale, heating, ventilation, and air conditioning (HVAC) units with AI technology are being installed in Canadian buildings. These “BrainBox” systems are able to raise, or lower temperatures based on inputted data such as weather forecasts, utility prices, and carbon-emission calculations. The producer estimate that these have cut energy emissions by 25%.
Here in the UK, we’re relying more heavily on Artificial Intelligence. The UK government has recently announced a huge £1.5 million programme encouraging developers to create programmes and applications that will cut the country’s carbon footprint.
The first stream of this project will be used to co-fund a centre of excellence on AI innovation and decarbonisation, and future fundings will be made available to projects that use AI as a tool to encourage a faster transition to renewable energy, change current industry practices and move them away from traditional carbon-emitting energy sources, and decrease agricultural emissions, one of the largest polluting sectors in the country.
So, in the UK at least, a greener future is dependent on AI. But what about the environmental cost of this technology?
Artificial Intelligence and Carbon Emissions
AI and applied machine technologies learning may be being used as a method to tackle climate change, but they’re not without their own carbon footprint.
CFR report that the supercomputers that are required to power modern AI technology produce up to, if not over, 250,000 pounds of carbon dioxide – per computer. These technologies are currently powered using the grid, so a switch to a renewable energy source in the future could see this number dramatically decline.
Furthermore, many smart devices, including those that run AI and more domestic products such as phones and tablet, have batteries built using lithium. As CFR report, lithium extraction is water-intense. Every tonne of lithium extraction requires approximately 500,000 gallons of water typically sourced from wetlands and water resources that don’t naturally recover. The result is a decline in the quality and volume of habitat for the existing wildlife.
Lithium may be labelled as “clean” in comparison to other energy sources, but there are still huge environmental implications.
There’s no doubt that our modern practices need to change if we are to secure the survival of our environment for future generations. Great strides have clearly been made in AI to enable this to happen, but there’s still steps to be taken to ensure these technologies don’t do more harm than good.
What are your thoughts on AI and Machine Learning Technologies as climate change resources? We’d love to hear your thoughts – get in contact to share your point of view!