"Right now is the youngest you will ever be." I was blown away by the profundity of the statement when I heard it first. Drawing parallels, a similar sentiment was shared recently, predicting the current summer might be the coldest we'll experience in future cycles. Of course, the non-El Nino summers will be better in the coming years (hopefully). But they do not mitigate the terrifying impact of what could be in store for the decades ahead.
I read The Heat Will Kill You First recently (https://bit.ly/3OtOIn0). For any global warming believers, there is little new information. Still, its anecdotal tales of the damage wrought by heat will chill anyone to the bone, particularly if they are inured to the news of wildfires, draughts, and floods.
Notwithstanding the contentious debates about ChatGPT and its siblings experiencing a cognitive decline lately, we can safely claim that right now, AI is the dumbest it will ever be. Generative AIs hold tremendous promise in tackling some environmental damage issues. Before one looks at the multiple positive ways, it is essential to scrutinize the AI-cranking data centers' carbon footprint. Research papers like this, https://bit.ly/3Kcnfni, have detailed that the carbon emission numbers of foundation models are relatively small overall and hence are outweighed by their potential benefits. Not everyone is as sanguine (https://bit.ly/3rMF1HG), and the AI builders will come under continuous pressure to improve the models' efficiencies in the years to come.
- Environmental science, a tapestry of intricate relationships and complex systems, could greatly benefit from AI's analytical prowess. For instance:
GenAI could revolutionize design processes, drastically reducing material usage without compromising functionality (https://bit.ly/3Y6To5s)
- Transformer models are best suited to study ill-understood, complex topics like fungal ecology and mycelium networks. An article published a few days ago shows how AI could prove invaluable in planning conservation efforts to protect the fungal network and identify fungi that can degrade plastics. https://bit.ly/43GtCqa
- Start-ups are trying to improve recycled materials' sorting, improve crop yields, or reduce manufacturing/building emissions using the enhanced abilities of the latest AI models. https://bit.ly/472zUDk
- The biggest direct benefits are from the impact on weather prediction in extreme weather events - https://ibm.co/3Qbu5xi. Benefits emerge in post-event salvage efforts, including firefighting and rescue.
Addressing climate change is a marathon involving countless small steps rather than a handful of grand gestures. GenAI could intervene in innumerable efficiency-enhancing ways as it is deployed in real-life practices. The keys to harnessing its power lie in our resolve to solve climate change and recognize AI's potential in untangling the Gordian knot of environmental issues.
Today is the latest we have to intensify our efforts.
