By Naomi M, MehtA+ AI/Machine Learning Research Bootcamp student
(Note: This plan is all hypothetical. We currently do not have the technology for terraforming the Great Basin. The Great Basin is located in Western United States. Terraforming is the process of making a place habitable for humans. By terraforming a land area and planting trees, we can absorb CO2 and potentially neutralize our CO2 emissions. )
In an age where the ability to find information with just a couple of clicks has been around for quite a while now, we have a new technology we are already embracing. Something that not only tells us information, leaving us to connect the dots with other things we know but connects the dots for us. Explains, plans, and engages. More than 1,800 million metric tons of Co2 were emitted in 2021 just from vehicles in the U.S. Why couldn’t ChatGPT find a precise solution to the large number of Co2 emissions from the vehicles we drive and need to transport goods daily?
Let’s Walk Through the Process
I started with a vague idea and started to narrow my questions down so that I could make sure all the details were figured out. By analyzing the answers of ChatGPT, we can improve the query which allows us to get more specific with our questions, eventually leading to the best answer possible.
Now this answer from ChatGPT is nothing we haven’t expected. It is general and doesn’t exactly provide us with the most actionable items. For example, ChatGPT recommends implementing irrigation systems and managing water resources, but we are in the desert. There is no water in the desert — or at least not enough to turn a desert into a forest — so we have to ask a more specific question about how ChatGPT meant this.
I was surprised when ChatGPT didn’t talk about desalinization as a way to increase the water amount in the Great Basin. So, after my initial question, I asked ChatGPT about it directly and it then began to explain the factors to be considered when using desalinization as a water source in terraforming the Great Basin.
Now we have our ‘how’ in getting the water needed for terraforming the Great Basin. Now I asked the amount of Co2 we want our new terraformed ecosystem to absorb. (How much is produced by vehicles in the U.S. per year). This number will later be used to find out how many trees we need to neutralize the Co2 emissions from transportation.
Knowing our terraformed ecosystem needs to intake a lot of Co2, we need to find out which trees intake the most Co2.
We don’t want the main tree in our ecosystem to die since it is not used to the weather and heat of a desert, so we can ask ChatGPT to pick the one best suited for high temperatures.
Since we have all the information necessary to make a more specific query for ChatGPT to answer, we just put all the data together in a question, letting ChatGPT figure out all the details, and what we would have to do.
Findings:
Something I noticed with ChatGPT, is that its response format is similar to similar types of questions. For example, in response to questions about how to or what would the plan be to terraform the Great Basin, it would respond in a list, with the first item being water management, and then soil or vegetation.
In terms of whether or not this plan could be put into use, it is important to note that current technology does not allow us to do this, however, while this is hypothetical, in the future, this could be used as a rough plan for terraforming the Great Basin.
While I expected the specifics of the cost of the project and the time the project would take to be vague, I didn’t expect the large number of actionable items given from ChatGPT, which when we later narrowed down the question, seemed pretty specific. I was especially impressed when ChatGPT could say the best trees for climate and Co2 absorption since knowledge of the specifics of trees is not what exactly comes into mind when first thinking about ChatGPT. In just the right question, ChatGPT gave an actionable plan to terraform a large desert. The applications of AI and ML are endless. I am very hopeful for the future of ML and know that ChatGPT has revolutionized the tech industry.
