Thank you, Mr. Chair.
I think this is an important discussion. I'm sure it will spill over into our next meeting.
I thank the NDP for bringing up an important amendment.
It's very important that we get the definition of “anonymize” correct. We had a lot of witnesses talk about anonymization versus de-identification.
I'm going off of what we were talking about before. The conversation was about the balance between what businesses need in order to interpret the risk and what the body responsible for enforcing the laws in the Privacy Act needs, which is the Privacy Commissioner, the OPC. They have to see certainty, not uncertainty, when they are interpreting the law for privacy. Based on our last amendment, we rely on the Privacy Commissioner to give us the best definitions to allow them to uphold the Privacy Act. The Privacy Commissioner and the tribunal will need those definitions to be exact.
To give you an example, the Privacy Commissioner, Philippe Dufresne, laid out a case during his appearance before the committee in our fall meeting on October 19, 2023. He said, “The bill says that more can be done with de‑identified information, and that if it's anonymized, the law doesn't apply at all. So there's a big responsibility that comes with that. These definitions need to be strict.” That's about how they interpret those definitions in the bill. I think that's why we're spending a lot of time here. Even though it's a preamble and a definition, it sets the tone for the rest of the bill and the rest of the conversations we're going to have.
Proposed subsection 6(5) of the CPPA exempts anonymized information from all protections it establishes. Given that our argument from this side—which you'll hear more than once—is that privacy is a fundamental right, we're making sure we get the definitions right when we go through this.
The Office of the Privacy Commissioner's submission from May 2023 talked about this piece and noted:
A final point relates to the new definition proposed for anonymized information. As currently drafted, organizations could anonymize personal information using “generally accepted best practices”. However, there is no explanation of what these practices are or what would be considered “generally accepted.” Including this language opens the door to the possibility that some organizations might rely on anonymization techniques promoted by certain experts or groups that are insufficient for a given dataset.
It could be insufficient for a given test regarding what we're trying to define and enforce under this privacy law. That is why it's very important. We have a problem, as the Privacy Commissioner did, with including the language “generally accepted best practices”.
Mr. Schaan, I know Mr. Turnbull just asked some questions on this, but do we have an actual legal definition of “generally accepted best practices” or a list that will be very specific for the OPC?