Federal government plans to increase its use of AI — with some big exceptions

Treasury Board President Anita Anand says she wants to increase the use of artificial intelligence in Canada's public service but insists there's no question of using AI tools like ChatGPT for confidential information like cabinet documents.

Speaking to reporters Monday as she launched consultations on the use of AI in the public service, Anand said the government is looking for a strategy to use AI responsibly.

"This is in the area of generative AI, in particular incorporating platforms like ChatGPT into the workplace," she said.

"Managers will have the discretion as to when that is able to be used, but I will say that in areas of confidentiality, cabinet confidences for example, generative AI isn't generally going to be used."

When asked why the government doesn't plan to allow public servants to use AI tools with confidential documents, Anand said that's part of a separate consultation.

"The issues relating to confidentiality are being addressed in our review of the use of confidential information across government," she said. "And so, I'm currently looking at that under a trust and transparency approach and I will have more to say on that specific issue very shortly. Not today, but shortly."

Anand said she wants the government's strategy to avoid bias or discrimination in the use of AI.

"This is the part of the work that we're doing to ensure that the AI tools that we are incorporating are not used for unsavoury or discriminatory purposes," she said. "It is part of the work that we're doing to use AI responsibly."

Anand acknowledged the government will have to compete with the private sector for the technological talent it will need to increase its use of AI but said there are plans to retrain existing public servants.

"We will be setting up a specific division of Treasury Board to re-skill existing public servants, if they would like to do that, in the area of cyber and cyber technology," she said.

WATCH: Open AI shows off model's capacity for conversation  

OpenAI demonstrates new model's capability for realistic conversation

7 months ago

Duration 0:42

From giving advice and analyzing graphs to guiding someone through a math equation and even cracking a joke, the new model of ChatGPT, called GPT-4o, is touted as having real-time responses in natural human tone.

Speaking to a hand-picked roundtable of experts Monday, Anand said the government has set three priorities for a national public service AI strategy: providing services to Canadians, increasing operational efficiency and driving science and research.

"The question is really going to be how can we, with this strategy, add greater efficiencies but also simplify the interactions of the Canadian population and organizations and businesses with federal services," Anand told the roundtable.

Anand said that while it's important to help the private sector develop AI, it's also important to develop the public service's use of AI.

"As the largest employer in the country, we set an example for those employers," she said.

Earlier, academics and experts participating in the roundtable stressed the importance of public trust in the government's use of AI, comparing it to the need for the public to trust the government's oversight of Canada's food safety system.

Officials told the roundtable that the government plans to launch its public service AI strategy by March.