Imagine that you work for FEMA and you are developing an artificial intelligence system to allocate emergency supplies in response to a hurricane emergency. (Do not assume the singularity ). You are in charge of ensuring that this system does not do anything crazy. Your system has to identify the supplies and the need for those supplies and figure out who gets what. It is complex. You have to send vital supplies to places in need, not send supplies to places that are not in need and prevent supplies from going to places that say they are in need but really are not. Your system also has to trigger resupply toyour storage facilities. Some at the extremes are easy. Some in the middle are hard. You have access to a lot of history about past hurricanes and their response. You are not inventing the system, your job is to make sure it doesn’t do anything crazy.

How can you use the history?

What safeguards would you include to ensure that everyone gets what they need?

Are there any warnings for the humans who are going to use your system?

At what point would you bring in humans to make decisions?

How could you eventually eliminate the need for humans in processing this paperwork?

In 500 words Describe the problem, describe your technology in brief (this is not about the invention), and then explore how you will keep this safe.