It would have been nice if the paper had included some description of the algos, like what factors they consider and how weighted, and the source(s) of the information. I wonder if clever defendents or lawyers would learn how to game the algos.
I am not questioning your post, as it is true that, in many scenarios, technology has an edge. However, there is another perspective that you may want to look at:
Based on my experience, both sides have some valid points we should consider before mass-deploying technology. As technology is very data-dependent, the real world deals with too many edge cases, and we do not have enough data available for every possible scenario/edge case. How do we keep the human in the loop is not easy too, as we start trusting technology blindly, which creates its own issues.
Yes, humans make more mistakes and have biases, but technology has its issues.
Unfortunately, we do not have a good solution to this problem, and we will need case-by-case solutions. This means that in some cases, technology will have an edge, in other cases, humans will have an edge, and in several cases, it will be something hybrid where workflow will require both to work together.
This is a form of technical support that should be welcomed by politicians and civil servants alike. It shouldn't replace these groups (beware of AI), but assisté them in making better decisions, as well as in defending those decisions.
It would have been nice if the paper had included some description of the algos, like what factors they consider and how weighted, and the source(s) of the information. I wonder if clever defendents or lawyers would learn how to game the algos.
Interesting study!
I am not questioning your post, as it is true that, in many scenarios, technology has an edge. However, there is another perspective that you may want to look at:
https://www.newyorker.com/culture/open-questions/in-the-age-of-ai-what-makes-people-unique
Based on my experience, both sides have some valid points we should consider before mass-deploying technology. As technology is very data-dependent, the real world deals with too many edge cases, and we do not have enough data available for every possible scenario/edge case. How do we keep the human in the loop is not easy too, as we start trusting technology blindly, which creates its own issues.
Yes, humans make more mistakes and have biases, but technology has its issues.
Unfortunately, we do not have a good solution to this problem, and we will need case-by-case solutions. This means that in some cases, technology will have an edge, in other cases, humans will have an edge, and in several cases, it will be something hybrid where workflow will require both to work together.