Some tasks seem far too difficult to tackle. For example, would you want to create a map of every tree in a forest and then use that in a model of some sort to evaluate how fire might spread? It sounds like something a well endowed foundation might tackle with the Amazon Mechanical Turk
and hundreds (or thousands) of humans.
However, with AI and satellites, this is a task that can be handled at a much lower cost. In fact, there's a company that is actually doing this
to help firefighters better do their jobs. With some overwhelming fires around the world in the last few years, this is something that those people on the ground will appreciate.
The power of computers and lots of data is becoming more and more valuable in all sorts of situations. Machine Learning and Artificial Intelligence algorithms can be especially powerful when working with lots of images, handling repetitive tasks better than humans at scale. Humans might do better with a particular image, but when there are thousands to review, they make too make mistakes.
The models used in AI/ML need to be checked and trained regularly, updated with new data sets, but the cheap computing resources can tackled problems that we find far too difficult to manage for most groups of people. They tasks are too mundane and repetitive and the scale too large to be cost effective for people to tackle themselves.
I expect to see more and more of this over time, where we use AI/ML to assist humans. People still need to define the problem, build models, and check their work (regularly), but use the computer to tackle the problem at scale. Humans also need to constantly evaluate if we are asking the right question or interpreting the data in a way that reflects the complexity of the world. However, that's the kind of partnership that will help us decide how best to manage the large scale world that we all need to live in together.