In a tax context, there is information that CRA obtains directly from the taxpayer. It is possible that the CRA looks at social media and other environmental information to gather intelligence and may put all of this information towards artificial intelligence. It's important. Data analytics is a new reality and it has many advantages.
However, AI systems need to be implemented in a way such that the information that feeds the system is reliable and has been lawfully obtained, so that leads to certain consequences. If CRA looks at information on social media, and let's assume for a second that it is truly publicly available, that says nothing about the reliability of the information.
To answer your question, in an AI context, privacy by design ensures that AI is implemented in such a way that the information that feeds the system, first, has been lawfully obtained, second, is reliable, and third, does not discriminate on the basis of prohibited grounds of discrimination, but is based on objective factors of analysis.