Right. Our concern is that without allowing for the definition of what datasets are on a yearly basis.... It is concerning. We believe there should be more clarity on what those datasets could and should be.
As well, we're worried, just as others have pointed out, that throughout the process of collecting those datasets has a changing threshold for what can be collected, what can be retained, and then what can be queried. Instead of allowing for so much information to be collected at the beginning and then narrowing it as we go, there should be strong requirements from the very beginning on of what information can go into those datasets.
Of course, as we've already mentioned, the fact that publicly available information could form a kind of dataset is a concern.
There is also concern that, at the very tail end of CSIS datasets, information collected in Canadian datasets can be accessed at a lower threshold for foreign surveillance purposes and that information collected by CSIS for foreign purposes can be collected, albeit at a higher threshold, from those datasets for domestic intelligence purposes. We believe that to fix that problem we need stronger authorization and stronger thresholds from the beginning. As we've mentioned, we also need an increase in the transparency and powers of the intelligence commissioner to verify those datasets and authorizations.