Imagine receiving a robocall. But instead of real people, that's the voice of a political leader who is telling you not to vote for you. You share it with your friends, your family – just to know that it is a surreal AI voice clone. This is not a hypothesis.
In January 2024, a fake Joe Biden Robocall arrived at New Hampshire Democrats, urging them to “stay home” ahead of the state primary. The voice may have been a composite, but the panic was real. And here's a preview of the threats faced by democracies around the world as elections become the most valuable target of AI-driven disinformation.
AI-generated content is becoming shockingly easy to create, with deepfakes, synthetic voices, artificial images, and more, and almost impossible to detect.
Unchecked, the harm brought about by this new threat of disinformation is countless, which could invade public confidence in the political system, reduce voter turnout and destabilize democratic institutions. Canada is not immune.
Read more: Deepfake use can soothe doubt and create confusion and mistrust in viewers
The danger is already here
Deepfake is an artificially generated media (video, audio, or image) that uses AI to impersonate real people. Benign applications (films, education) are well understood, but malicious applications are quickly catching up.
Open source generation AI tools such as ElevenLabs and Openai's voice engine can generate high-quality cloned voices with seconds of audio. Apps like Synthesia and Deepfacelab perform video operations in the hands of those who have a laptop.
(AP Photo/Caroline Custer)
These tools have already been weaponized. Beyond Biden Robocall, Trump's campaign shared an AI-generated image that Taylor Swift supports him.
Meanwhile, state legislative entities are rolling out deepfakes of coordinated disinformation campaigns aimed at democracy, according to the Knight First Amendment Institute, a free speech advocacy organization.
Why it matters to Canada
Canada recently concluded a federal election in 2025. This was implemented without strong legal protection against AI-enabled disinformation.
Unlike the European Union, where AI laws are enacted that require clear labeling of AI-generated text, images and videos, Canada does not have binding regulations that require transparency in political advertising or synthetic media.
Instead, they rely on voluntary codes of conduct and platform-based moderation, both of which have proven inconsistent. This regulatory gap makes Canadian information ecosystems vulnerable to manipulation, especially in the context of minority governments where another election could be called at any time.
(AP Photo/Michaeldwire)
Alarms are installed all over the world. A survey by the Pew Research Center in September 2024 found that 57% of Americans were worried that AI would be used to generate fake election information. Canadian polls show similar levels of concern.
Researchers near their home recently discovered a Deep Fark clip that mimics CBC and CTV Breaking News. This is getting caught up in preparation for Canada's 2025 vote.
What we can do
A single solution is not a panacea, but Canada can take the next important step.
-
Content Label Law: Emulates the European Union and delegates AI-generated labels for political media. The EU requires content creators to label manufactured content.
-
Detection Tool: Invest in deepfark detection research and development in Canada. Some Canadian researchers are already doing this, and the resulting tools need to be integrated into platforms, newsrooms, and fact-checking systems.
-
Media Literacy: Expanding public programs to teach you how to find AI literacy and deepfakes.
-
Election safeguards: Equipped Canadian elections with rapid response guidance for AI-driven disinformation.
-
Platform Accountability: Keep a platform responsible for not being able to act on a validated deepfake basis, and requires AI to provide transparent reporting on how to delete and detect content generated by AI.
Empower voters in the AI era
Democracy is built on trust in the information consumed by elected officials, institutions, and voters. If they cannot trust what they read or hear, that trust will erode and the very structure of civil society begins to unravel.
AI is also part of the solution. Researchers are working on a digital watermark scheme to track manufactured content and media outlets, and deploy fact-checking using machine learning. Staying ahead of AI power disinformation requires both smart regulation and alerting masses.
The political future of Canada's minority governments is uncertain. We cannot wait for the crisis to take action. By modernizing the law and building a proactive infrastructure, taking action now helps ensure that democracy is not another victim of the AI age.
