Consider data on train positions in the UK. It's been locked up in various ways for a long time, or at least with restrictions that amount to non-commercial usage. So what you ask? Providing a live train-times service from the data costs money: I've been told it's upwards of 500 messages/second and there's a lot of problems you have to work around.
Releasing the data behind it is starting to help people to find cheaper tickets, make a more truthful record of train punctuality than official figures, etc. Many companies will benefit financially from this (rather than the current monopoly) but the public will benefit.
There are cases where accountability is at least starting to be possible. Last year I helped a newspaper look at published NHS data for things buried under the carpet - think hospitals that are failing to treat certain age groups for conditions like Diabetes. That's accountability. The UK pupil database is similarly promising.
Things need to get better but it's not as bad everywhere as the author claims of Canada or the USA.