The Canadian government, via its Treasury Board, now has a month old Directive on Automated Decision Making. The directive, applying to federal government departments, is to guide the application of Automated Decision Systems. The intent is to ensure the use of such systems is done with limited risk, with procedural fairness, due process, with assessment of impacts and outcomes and with information made to the public.
An Automated Decision System as defined in the Directive is:
any technology that either assists or replaces the judgement of human decision-makers. These systems draw from fields like statistics, linguistics, and computer science, and use techniques such as rules-based systems, regression, predictive analytics, machine learning, deep learning, and neural nets
Impact assessments must be made in the use of Automated Decision Systems. Notice of their use and how a decision is made is required. A series of quality assurance provisions are included. So far so good, and relatively banal. But then tucked into the definitions in the Directive is this:
Information technology that performs tasks that would ordinarily require biological brainpower to accomplish, such as making sense of spoken language, learning behaviours, or solving problems.
[See the end of this post for the definition of Infomation technology.]
Remarkably, the Directive doesn’t actually seem to use the term “artificial intelligence” except in the descriptive opening paragraph about the Directive, which reads “[t]he Government of Canada is increasingly looking to utilize artificial intelligence to make, or assist in making, administrative decisions to improve service delivery…” Nor can I find this definition used anywhere else with a web search. Did the Canadian federal government just sneak out an (almost) legal definition of artificial intelligence?
The EU has a recent discussion paper on the definition of artificial intelligence that came out at about the same time. At the pdf linked to that site you can find the discussion which ends with a proposed definition. But as far as I know they have not rolled that proposed definition into any legislative effort. There is no clue I can find as to where the Canadian Treasury Board got its definition.
Nor can I find any other government that has gone this far.
So, given that the Treasury Board acts as the delegate of the Queen’s Privy Council and the Governor in Council, does this mean that the Canadian Treasury Board is the first government entity to roll a definition of artificial intelligence into a document that has legal authority?
Information technology is defined in the government’s directive on Management of Information Technology and is defined as : “”information technology (technologies de l’information)Involves both technology infrastructure and IT applications. Technology infrastructure includes any equipment or system that is used in the automatic acquisition, storage, manipulation, management, movement, control, display, switching, interchange, transmission or reception of data or information. IT applications include all matters concerned with the design, development, installation and implementation of information systems and applications to meet business requirements.”
In a subsequent post I’ll compare the Canadian and EU definitions.
3 thoughts on “Did Canada roll out the first legal defintion of artificial intelligence?”