Yes. For the past 14 months we've been working on direction and policy for the way government departments use automated services, moving forward, because this is something we're seeing increase.
You heard me say that PSPC has done a procurement action for creating an AI vendors catalogue. We now have 74 companies that have been registered in this vendors catalogue for the the last few weeks. We've been providing policy direction, working fully in the open, putting our policy instruments and our documents and our directives on GitHub and Google Docs for the world to see. We've collaborated with MIT and Oxford and other universities.
It's been a very open and transparent process in creating a tool set that is called an algorithmic impact assessment. You would be familiar with something called a privacy impact assessment, when we're looking at privacy; we're looking to develop something similar, as a prevention tool, for algorithms that will identify to the departments the level of risk. Automating a chatbot that tells you through the NCC website whether the skating rink is open is probably a lower-risk step than automating our borders, for example, or our immigration systems.
We've been developing this tool fully in the open. We've had countries such as Mexico start to use the algorithmic impact assessment already, and we have other countries, such as Portugal and others we're working together with—the “digital nine”—use some of the tools we've been developing. This creates an international body of knowledge that up until now just did not exist.