The Communications Security Establishment released a report on the cyber threats to Canada's democratic process.MARK BLINCH/Reuters
Canada’s cybersecurity protection agency is warning that China, Russia and Iran will “very likely” use artificial intelligence in an attempt to interfere with the 2025 federal election expected this spring.
The Communications Security Establishment released a report Thursday, the Cyber Threats to Canada’s Democratic Process: 2025 Update, that details the impact increasingly powerful and accessible artificial intelligence tools have had on efforts by foreign actors to spread disinformation, harass politicians and conduct espionage or interference.
These include the means to make deepfake photos, videos or audio, or write computer code. AI tools rely on machine learning, including “large language models” trained on big sets of language data that can be used to write personalized e-mails designed to gain access to a target’s computers or personal details.
CSE said China, Russia and Iran will likely use AI tools to spread disinformation and mount “hack and leak operations” designed to influence target audiences through the disclosure of privately-held material.
“Canadian politicians and political parties are at heightened risk of being targeted by cyber threat actors, particularly through phishing attempts,” it said in a statement, referring to efforts by hackers using e-mails or texts to con people into revealing sensitive information or installing malware on their computers.
Hogue inquiry finds some parliamentarians’ conduct troubling, but not treasonous
The CSE says “generative AI,” or artificial intelligence that can create or alter photos and videos and write text or computer code, used to be expensive and required deep technical knowledge to use. But the advent of improved AI models and user-friendly web interfaces, combined with few regulations or guardrails, has made it relatively easy for threat actors to engage in malicious activity, the agency said.
The goals of China, Russia and Iran are “almost certainly” to break alliances between democratic countries, “entrench divisions within and between democratic states while advancing their own geopolitical goals.”
It said these foreign actors are using AI to create disinformation and “flood the information environment with false information,” including by relying on botnets – hijacked computers infected with malware and viruses – to spread this material.
Canadian public figures, especially women and members of the 2SLGBTQI+ community, are at heightened risk of being targeted by deepfake pornography. “Without updated legal and regulatory guidelines, we assess it very likely that the spread of this content will continue unabated,” it said. “Artificial intelligence is being used in this way to humiliate, intimidate and exclude targeted persons from political participation.”
CSE said Russia’s disinformation campaigns do not gain significant visibility “without the amplification of witting or unwitting actors” from within a targeted country. A 2024 deepfake disinformation campaign targeting Democratic vice-presidential candidate Tim Walz, which falsely alleged he abused someone while working as a high school teacher, only gained significant attention after it was covered by right-wing American commentators.
China is likely to employ AI-driven interference to “push narratives favourable to its interests and spread disinformation among Canadian voters,” CSE said. This includes targeting Chinese-diaspora communities in Canada.
How scammers are using AI to create convincing deepfakes, and authorities are using it to catch them
As The Globe has previously reported, China wants to build acceptance abroad for its claims on Taiwan, a self-ruled island it considers a breakaway province and reserves the right to annex by force. And it seeks to play down its abuses in Xinjiang, where the office of former UN Human Rights commissioner Michelle Bachelet in 2022 said China has committed “serious human-rights violations,” which may amount to crimes against humanity.
Similarly it wants to generate support for a draconian 2020 national-security law to silence opposition and dissent in Hong Kong, a former British colony Beijing had once promised would be allowed to retain Western-style civil liberties for 50 years.
Beijing also seeks to quell foreign support for Tibet, a region China invaded and annexed more than 70 years ago, and to discourage opposition to Beijing’s militarization of the South China Sea and its sweeping maritime claims in the region.
While Canada’s general elections are conducted by paper ballot, the CSE noted much of the “electoral infrastructure” in this country, including voter registration systems, election websites and communications between election management bodies and their employees, is digitized. “This creates a threat surface vulnerable to malicious cyber activity aimed at compromising the confidentiality, integrity or availability of the underlying system before or during an election period,” it said.
“Despite these threats, we judge that it is very unlikely that AI-enabled activities will fundamentally undermine the integrity of Canada’s next general election,” the CSE said.