Ottawa is developing a new artificial intelligence strategy for federal government operations, Treasury Board President Anita Anand said on May 27.
“It is a big step towards a cohesive and consistent federal approach to AI,” Ms. Anand said at a meeting of AI experts in Gatineau, Que., where the strategy is being discussed.
The intention is to make government operations more efficient and improve services for Canadians, while also improving capacity for science and research.
Anand said the question is about how to “add greater efficiencies but also simplify the interactions of the Canadian population and organizations and businesses with federal services.”
She said AI could be used to automate routine tasks, but said generative AI “isn’t generally going to be used” when it comes to confidential uses like those involving containing cabinet confidences.
The effort is not about reducing the number of jobs in the federal public service, she said.
The initiative will include setting up a specific division to retrain existing public servants. Ms. Anand said the government will also work on hiring “quickly and competitively” to bring in more top tech talent.
In addition to Monday’s meeting of experts, the government will consult broadly in the coming months before unveiling the strategy. There is no deadline for producing the strategy.
The federal government has already begun incorporating AI into its operations, using the technology in hundreds of different ways.
As the largest employer in Canada, the federal government sets an example for employers in the private sector, Ms. Anand said.
She said there is a need for balance when it comes to responsibly using AI, which includes ensuring that privacy laws are respected and that AI isn’t used for “discriminatory or unsavoury” purposes.
The announcement came as eight more private organizations signed on to a voluntary federal code of conduct for artificial intelligence, including Mastercard, Lenovo and the Quebec real-estate brokers’ association.
Adopters of the code launched last year must agree to undertake several measures aimed at reducing the risks of AI, including screening data sets for potential biases and monitoring systems for possible harms.
