It's very nice of AI to “clean up after itself." 

As governments, industry, and individuals worldwide look to reduce their carbon footprint, the emergence of AI and the computing power it requires (and therefore the energy it consumes) is offsetting some of the gains made. 

According to this article from the World Economic Forum:

• As AI accelerates in importance to people and the economy, its significant energy demand and consequent environmental impact is also drawing attention.
• Integrating AI into renewable energy generation — particularly solar power — could improve efficiency to offset the tech's demands on the power grid and associated emissions.
• Already, use cases like predictive maintenance and AI-enabled trading are emerging as ways AI can improve the deployment of solar power.

Maximizing the efficiency of solar power – which includes both capturing/maintaining the energy as well as understanding how and when to deploy it – is something that has eluded the industry for decades. It's getting better but seemingly has a ways to go. But with AI enabled energy systems, we might at least be able to solve the deployment piece of the puzzle. 

And if that can offset the impact of the additional power consumption, we can get back to taking a bite out of the rest of the carbon creators.