I think that most of us have already heard about Hurricane Ian and Fiona. Both of them occurred in September 2022 in North America and Latin America. They did a lot of damage, especially to Cuba, the southeast United States, Canada, Puerto Rico, and the Dominican Republic.
To help people who suffered from hurricanes, the nonprofit GiveDirectly collaborated with Google. Thanks to the latter, GiveDirectly knew by provided images which areas were the most destroyed and needed financial support. A Google tool called Skai was used in that process. It analysed satellite imagery from before and after a disaster and estimated the severity of damage to buildings. After getting to know that information, people in need were given a one-time payment of a total of 700$ in order to buy the most necessary things like for example food.
In my opinion, this kind of activity of AI led to effectiveness. People who needed money did not have to queue for handouts in public because GiveDirectly already knew to whom they should have given money thanks to the help of Google. It made the process of the selection of whom to aid faster. On the other hand, this kind of solution was also dangerous because AI is not a human and it could not check indeed if the chosen person seriously needed the help. It is not like the help of aid workers who can talk with the people and notice if the person needs financial support. For example, let’s say that the person has one of the most damaged property and Skai detected it. That’s how that person, later on, gets the money. What if he or she is rich and does not need the money? Does AI detect it? No. Of course that the person can return the money, but firstly it would take time, and secondly some people are greedy, so they would not tell anyone about the situation. At that time, people who need the money do not get it and it can be crucial for them because of for example hunger.
Thanks for your time. Feel free to comment below. 🙂
References:
https://en.wikipedia.org/wiki/Hurricane_Ian#Aftermath
https://en.wikipedia.org/wiki/Hurricane_Fiona#Aftermath
https://screenshot-media.com/technology/ai/artificial-intelligence-hurricane-relief/
https://www.independent.co.uk/tech/hurricane-ian-damage-relief-ai-b2199411.html
Very interesting topic but I see a few disadvantages, first is the same as you maintained in the blog second is that this program should support people responsible for spotting people in need after a hurricane and should not choose them himself but under the supervisor.