Canadians and Canadian businesses have until Sept. 23 to tell Ottawa how artificial intelligence systems should disclose what they are and what they can do, in a federal consultation that has drawn comparatively little public attention since it opened in July.
The consultation was launched July 23 by Evan Solomon, Minister of Artificial Intelligence and Digital Innovation, according to a news release from Innovation, Science and Economic Development Canada. It seeks views in five areas: detecting and identifying AI-generated content, letting people know when they are dealing with an AI system rather than a person, standardizing the information companies publish about how their systems were built and where they fail, tracking serious AI-related incidents, and tracking what AI agents do when they act on a user’s behalf.
The sharpest case for the exercise sits inside the government’s own discussion paper rather than in the announcement. Citing a KPMG survey of large Canadian businesses, the paper reports that nearly three-quarters of respondents lost between one and five per cent of their annual profits to AI-powered fraud over the preceding year, and that deepfaked documents and voice impersonation calls of company executives were among the most common attacks they encountered. The paper also cites the Canadian Anti-Fraud Centre’s finding that scammers increasingly use generative AI tools, including deepfake videos of prominent people promoting fraudulent investments. In one 2023 case referenced in the paper, an Ottawa couple lost a significant sum after acting on such a video.
Businesses weighing a submission should note two conditions the department sets out on the consultation page. Submissions are treated as public documents and may be posted online, and participants are told not to include personal or confidential information. Responses can be filed through an anonymous survey or by email to the department’s AI consultations address.
The department also states that, depending on the volume of submissions, AI tools may be used to process the feedback received. Following the consultation, it says it will publish a What We Heard report, though it has not committed to a publication date or to specific legislation flowing from the results.
The transparency file is one strand of a broader push. Ottawa released its refreshed national strategy, AI for All, on June 4, and has since introduced Bill C-36, the Protecting Privacy and Consumer Data Act, and Bill C-34, the Safe Social Media Act, which proposes obligations for social media operators and some chatbots. On Sept. 3, the government released a set of data centre principles signed by major AI and cloud companies including Amazon, Anthropic, Google, Meta, Microsoft and OpenAI, a voluntary framework Solomon said does not create a new approval process and does not override provincial or municipal approvals.
That pattern matters for how much weight to give the current exercise. Several of the instruments Ottawa has reached for this year are voluntary codes rather than enforceable rules, and the discussion paper concedes as much about the existing Canadian code for AI developers, which recommends maintaining a database of post-deployment incidents but, in the paper’s words, is voluntary and not enforceable. The paper also notes that marking and labelling requirements were removed or streamlined from an early draft of the European Union’s Code of Practice after concerns about compliance burden, a signal of where similar pressure may land here.
For Canadian firms already running customer-facing chatbots, or agentic tools that can spend money on a user’s behalf, the days left in the window are the cheapest opportunity they will get to shape what disclosure eventually costs them.
Sourcing: Two ISED primary sources read in full, the July 23 news release on canada.ca and the “Have your say” consultation page on ised-isde.canada.ca, plus the full discussion paper “Enhancing trust in artificial intelligence through increased transparency.” Independently corroborated by Canada’s National Observer’s Sept. 3 coverage of the data centre policy.
