Thinking on past projects, the one that comes to mind right away is the app branding project we did in level 4, especially because mine ended up being a really interaction based thing.
Basically, my app was to encourage disposing of cigarettes properly, incentivising you with rewards and helped show you places you could get rid of them on a map. The map had locations and images to help show you where it was and looked like.
So, I think AR would be a good for this, for example making the locations really obvious through the phone, like having something there in the AR. Another way I can think of using it is to have codes at the locations, encouraging people to use the app and find the bins or whatever to get rid of their cigs because they'd have to actively look for the reward (my app had codes which you could earn which would give you things like discounts).
I really like that last idea for that project since it kinda forces the user to actually seek out a proper disposal. Obviously AR is probably a bit much for a basic little app and would be hard to implement, or maybe not, I don't know anything about that stuff.
AR is interesting, although I think, from what I've seen, the models look kinda simple. I also don't like you have to see everything through your phone for example. It seems people get use to it and the initial excitement goes away, like with Pokemon Go, that was very popular and now most people don't have the app anymore. I think it would be good for the individual user experience, to have your own view and things to find for you alone or to make things for you alone, but I don't really like that.
Future projects, I'm more drawn to the data driven design, just for the infinitely generated results from the same kind of data, so there would always be similarities in the results which is interesting. I see that as less individual, it comes from pre-existing ideas and work, and even though it might be instant like AR can be, it's more connected I guess.