Disasters.

Within the past week, there's been a jerk running over people in Nice, France, and a coup in Turkey.  Both of these are "I would rather not be there" kind of situations, which can easily be generalized to "some places become suddenly dangerous, and we should be able to help with technology."

This thought came to me this weekend while thinking about all of these events in combination with Pokemon Go, and then through the filter of "doesn't Twitter have a lot of information?"  I'm sure they have a similar level of information as PKMNGo, so it seems like they could set up a fairly robust "Emergency Mode."  Here's kind of a rough outline of things.

1. Twitter detects a disaster.  Although this could be a top level decision "Nice is in trouble.  Flag it!" it could also be something that you let be detected from the tweet stream.  It should immediately start trending in an unexpected way, which might trigger things.
2. A region of the globe where the disaster is happening is defined.  Every user in that region has the following things happen automatically.
1. Their tweets no longer contain location information.  This information is collected, but it is not released publicly.  There are many of these disaster scenarios where you have antagonists who might benefit from knowing where people are, so it's best to keep that private.
2. All users are presented with a pre-defined tweet screen.  They don't get to add to the tweets, but select from a menu:  "I am fine, and I am in a safe location."  "I am fine, but do not think my location is safe."  "I am not fine, and need emergency help."  These go out to the user's followers, so they don't need to use further communication resources contacting people.
3. From the known location information and the responses, narrow the disaster region to define safe and unsafe areas.  Share emergency information with authorities so they can move resources.  This can also be shared with users who do not think they're in a safe location.  "Shelter in place if you are secure.  If not, this map has been auto-generated, but shows the areas currently believed safe."
4. Pre-defined lists of "Emergency Users" are added to the users' following for a temporary period.  This might include news sources, but also things like the State Department's travel twitter.  I already follow them, and during the coup, they were spreading the information they had (largely "shelter in place").  I'm sure other countries have similar things, and can be added to lists for users from those countries (which twitter can easily sort out from location information, in addition to the user supplied location field).  The National Weather Service would be another option, for tornadoes, hurricanes, etc.
5. One option could be to add prompts for tweets/periscopes/etc. so that people who are in a place where reporters aren't could get information out.  This might add a lot of liability, though, because some people aren't smart enough to not put themselves into danger.
3. Once the disaster is resolved, everything goes back to normal, and the temporary following lists are deactivated.

I'm sure a similar thing could be set up with Google or Facebook, but since I use Twitter for pretty much all of my news, that's my obvious source.  Plus, they could make it a selling point: "even if you don't care in general, why not get Twitter so you can be on top of emergency situations."  I'm also pretty sure that news agencies would be interested in getting marked as an "Emergency User".  Fold in some after-the-fact "was this source useful?" polling, and they could check that those users actually help.