On June 18, 2026, Canada's Strong and Free Elections Act — Bill C-25 — received Royal Assent as Statutes of Canada 2026, chapter 20, the most substantial overhaul of the Canada Elections Act in a decade. The legislation targets three distinct pressure points: synthetic media impersonation of candidates, foreign-entity access to digital advertising platforms, and a penalty regime whose maximums had not kept meaningful pace with modern enforcement needs.
The context is empirical rather than hypothetical. A peer-reviewed study of the April 2025 Canadian federal election, presented at the ACM Web Conference 2026, found approximately 5.86 percent of all election-related images on X, Reddit, and Bluesky were AI-generated deepfakes. Credible-looking synthetic images attracted disproportionately high engagement relative to their overall volume. Canada's Public Inquiry into Foreign Interference in Federal Electoral Processes had already called for legislative fixes of precisely this kind.
What the Law Actually Does
Deepfakes (Section 480.1) — Bill C-25 criminalises creating or distributing a manipulated image or voice imitation that is likely to be mistaken for a candidate, party leader, or chief electoral officer when the intent is to mislead. Conduct that is "manifestly for the purpose of parody or satire" is explicitly exempt. The provision closes a genuine loophole: prior anti-impersonation rules required the creator to directly pose as the targeted individual. AI-generated proxy impersonations — where a third party generates a convincing fake without claiming to be the person themselves — fell outside that prohibition.
Foreign Platform Advertising — Broadcasters, newspapers, and online platforms are now prohibited from knowingly selling advertising space to foreign entities for election or partisan advertising purposes. The restriction applies year-round — not just during the formal election period. The previous regime contained a structural gap: foreign-linked operations could legally purchase partisan advertising on Canadian platforms throughout the inter-election months, constrained only once an election was called.
Penalty Escalation — Maximum administrative monetary penalties for individuals rise from $1,500 to $25,000, a roughly seventeen-fold increase. Corporate maximums go from $5,000 to $100,000. The Commissioner of Canada Elections gains expanded investigative tools, including the power to hold those who counsel violations to the same liability standard as those who commit them. Separately, $31.5 million over five years is directed to Global Affairs Canada's Rapid Response Mechanism for detecting foreign information operations.
The Case for These Measures
The strongest argument for C-25 is that Canada had already lived through the specific threats it addresses. The Slovak deepfake scandal of 2023 spread fabricated audio of a candidate discussing vote-rigging in the days before polling. A deepfaked Joe Biden robocall in New Hampshire's 2024 primary discouraged Democrats from participating. Canada's own 2025 election produced a measurable 7.9 percent synthetic image rate on X. Restricting foreign entities from buying year-round partisan advertising in Canadian digital space is the kind of proportionate, targeted response that evidence-based regulation should look like. The parody exception in the deepfakes provision is thoughtful drafting that protects satirists without shielding bad actors.
Where the Law Falls Short
The Senate's own committee offered a pointed warning. The Standing Senate Committee on Legal and Constitutional Affairs passed Bill C-25 without formal amendment but explicitly regretted that time pressure prevented "more extensive testimony on how the bill addresses foreign interference and developments in artificial intelligence, including the use of deep fakes." The chamber designed to provide sober second thought on novel technology policy was denied adequate time to exercise that function on the bill's most consequential provisions.
The Department of Justice's Charter statement acknowledges that the deepfakes ban and false-statements provisions engage section 2(b) expressive freedoms, defending the restrictions on the grounds that they target "lower value expression" that undermines elections. That framing is legally defensible but analytically thin — the lower-value doctrine developed in commercial speech and obscenity contexts, not in political speech. Its extension here received less parliamentary scrutiny than the stakes warrant.
The deeper problem is structural. Law professor Robert Diab, writing specifically on C-25's deepfakes provisions, argues the criminal ban "won't restore credible news to the platforms where many Canadians look for it." The point is sharper than it sounds: Meta's 2023 news ban created an information vacuum that deepfakes and foreign-linked content rushed to fill. Criminalising synthetic impersonation does nothing to restore journalism competition that makes manipulation less effective. The bill imposes no affirmative content detection obligation on platforms — it outlaws the behaviour without requiring anyone to find it.
The foreign ad ban carries a parallel enforcement problem. Platforms are prohibited from "knowingly" selling ad space to foreign entities — a scienter requirement that a sophisticated influence operation can evade by routing purchases through domestic intermediaries. The Commissioner of Canada Elections still depends on complaints and observable violations, not algorithmic monitoring.
Where Canada Now Stands
C-25 places Canada among a growing cohort of jurisdictions enacting election-specific deepfake prohibitions — over 45 US states have done the same at the sub-federal level. The European Union took a structurally different approach under the AI Act, mandating transparency labeling for AI-generated content rather than purely criminal prohibition. Canada's architecture is punitive rather than systemic: it bans without compelling disclosure or detection infrastructure.
The legislation is right on principle. Deepfakes designed to deceive voters about what candidates said or did are a genuine threat to informed electoral participation. Foreign entities funding partisan advertising in Canadian elections should face a clear prohibition. The penalty regime was embarrassingly weak at $1,500 per individual and needed meaningful escalation. All of that is true, and C-25 delivers on all of it.
What it cannot resolve is the gap between outlawing behavior and ensuring platforms detect it. A criminal prohibition enforced by complaint, against synthetic media that scales cheaply and routes through domestic buyers, is a serious starting point — not a finished solution.