There are a number of issues with the app some of which have been talked about in other articles and others have not.
One problem is that it is not clear if other types of data will be used to supplement the crime data.For example, an area might be labeled unsafe by the app simply because there are a lot of assaults in an area, but the number of assaults that occur in an area may simply be a function of the number of persons who reside in the area. This is the classic count vs. rate problem. At this point it is unclear how exactly neighborhoods will be classified. In order to calculate the crime rates other population data would also need to be integrated into the app.
Another issue is that the number of assaults may be influenced by some other environmental factors that can explain a large percentage of the crimes that occur in an area. For example, lets say a user has been alerted that an area is "BAD". We know from criminological research that areas that have large numbers of bars may have more assaults due to the fact that many assaults occur in places where young people and alcohol come together in space. If we control for the effect that bars have on the number of assaults, then a place that is labeled "BAD" may not be so scary to people who don't plan on going into bars. In other words, the new app may label a place as unsafe for a pedestrian, but not consider the fact that most assaults in the 'alerted' location are due to altercations that happen at bars. Therefore, as long as the person avoids bars the neighborhood is basically a safe place!
Another problem with the app that has many people upset is that labeling certain areas as "BAD" can have negative consequences on people perceptions of neighborhoods that are only based on Microsofts interpretation. Negative perceptions of neighborhoods can have cyclical effects on neighborhood economic performance. Fewer people will likely attempt to enter alerted neighborhoods which would negatively affect local business owners. Another important question is how does this affect the ability of homeowners in certain places to effectively sell their home? Labeling in most circumstances is not a positive thing to do, especially when the labeling involves large numbers of people.
As any technology advances there will be issues with ethics. Certain technologies may help some while simultaneously putting others at a disadvantage. The argument could be made that while it is true that the app labels certain neighborhoods, Microsofts app simply supplements the natural map that each of us carries as we learn and experience our environments. There are likely places that you choose not to go to either because you have heard that they are bad, or because they make you feel unsafe. Perhaps an app like this allows you to make those decisions with actual data instead of stereotypes and misinformation.
Check out other articles on Microsofts new GPS app at these sites.