AI and the energy crisis

It is going to take a lot to undo decades of carbon emissions. The solution can't be unilateral or focused solely on one solution, but some steps have a higher return than others. One of the tools expected to have the highest impact on reversing climate change is not recycling or renewable energy, or all becoming commune living vegans. 

It's AI.

According to the Capgemini Research Institute, AI is expected to aid companies across several industries to fulfil as much as 45% of the emission goals set forth by the Paris Climate Agreement by 2030. How much AI can help different sectors is, of course, determined by how many companies adopt AI in the coming years. If industries adopt multiple AI strategies, which has its own difficulties to be fair, the eventual return will be much higher than if we sit with the status quo. 

AI is so powerful not because of its strength but rather its flexibility. The way the automotive industry utilizes AI is wholly different than how the energy industry would. AI's malleability can help companies across various sectors reduce their carbon emissions in multiple ways. Even in the next three to five years, according to the Capgemini Research Institute, AI will likely reduce greenhouse gasses by 16% and improve power efficiency by 15%.

Energy 

Advancing renewable energy sources is a huge part of reducing carbon emissions, but AI can help make all energy forms more efficient. Machine learning can predict maintenance issues, prevent costly breaks, note leaks, and better aid distribution to the power grid between various power sources.

Another issue is anticipating the peaks and lulls of energy consumption. By predicting a building’s thermal energy needs, heating and cooling can be rationed out at optimized times. This benefits not only general power consumption, but can lower utility bills by taking out energy when it is the cheapest. 

Wind farms are one of the places we have seen vast improvements thanks to technological innovation. AI  can coordinate an entire field of turbines to act as one, minimizing "wasted" air flows and maximize power generation. One example of this is DeepMind, a subsidiary of Google, which developed a system that allows wind turbine facilities to predict wind energy production and power demand up to 36 hours in advance. This gives wind power plants the ability to plan energy production, reduce waste and provide a better price point to customers. 

Recycling 

We all like to pat ourselves on the back when we separate our recycling, but the reality is that because of one reason or another, a large percentage of what we throw into the recycling bin still ends up in landfills. The main culprit- us. We put things in the recycling bin that aren't recyclable, often by accident, sometimes by carelessness. All of this trash can break recycling facilities machinery and cause loads of recyclable material to be thrown out due to contamination. 

Until recently, the way to solve this problem was to have workers staggered across an assembly line, picking out all of those non-recyclable plastics and general garbage that get thrown in the bin—clearly neither a particularly effective nor affordable solution. By now, many companies have found ways to use AI to sort our recycling better. Many of them incorporate automatic sorting machines that can spot different kinds of materials faster and with far higher accuracy than the human eye. Essentially a giant claw game, darting back and forth cleaning out the garbage until only the recyclable materials are left. 

Automotive

Navigation applications help you optimize your route by avoiding tolls or getting home faster by avoiding accidents and construction. We utilize these features because they are useful and improve our journey. If your navigation system also helped you find a route that reduces your car's carbon emission, would you opt for that route? Commercial fleets already implement this kind of technology to optimize routes based on traffic conditions, signals, and shared mobility options, and it adds up. If we were to optimize every car on the road with the same tools, the combined effects would be substantial. 

Sensors 

This is a bit of a broader topic that could easily be its own book. There are so many use cases for different types of sensors that they could be considered the Rosetta Stone of AI in pollution reduction. Massive amounts of energy is wasted just because systems aren't efficiently set up. Leaks, breaks, or any number of minute small issues can add up quickly. 

According to many climate scientists, one of the offenders that is the easiest to change is the amount of energy waste in buildings. Most buildings are rather old and are almost a poster child for energy waste by this point. Anything from lights being left on, to temperature control or air leaks can slowly but surely pick away at the energy efficiency of a building. It surely wouldn't be cheap or easy, but retrofitting buildings with sensors to monitor and regulate the building's conditions could reduce energy use by 20%. Multiply that by every structure, that makes these upgrades, and it becomes a very significant number. 

Reforming and retrofitting the world’s power grids, structures and general culture won’t be easy, but at this point, no way out of our current environmental conundrum will be. AI might provide a real way out of the situation that we find ourselves in, but it will require quick, and possibly difficult action.