Canada’s elections chief warns AI could quietly influence voters and undermine democracy

By: Ebenezer Adugyamfi & Emmanuel Ayiku for GhanaianNewsCanada
November 3, 2026

Canada’s Chief Electoral Officer has warned that artificial intelligence (AI) is creating new risks for the country’s democratic institutions, cautioning that its influence on elections extends beyond fake videos and digitally manipulated images.

In remarks delivered before the House of Commons Standing Committee on Procedure and House Affairs on Tuesday, November 3, the head of Elections Canada said AI was changing how political information is created, distributed and consumed, making it increasingly difficult for voters to understand what information is genuine, who is behind it and why particular messages are reaching them.

The warning comes as Canadians increasingly turn to AI-powered chatbots and other digital tools for information about elections, candidates and political parties.

The Chief Electoral Officer said one of the greatest concerns is the lack of transparency surrounding how artificial intelligence systems influence the information environment.

He explained that AI can produce sophisticated content quickly and at relatively little cost, while also helping distribute and promote information to targeted audiences.

This means that both accurate information and false or misleading claims can be amplified to reach large numbers of people.

According to Elections Canada’s presentation, the challenge is no longer limited to identifying fake photographs or videos. It also involves understanding how AI systems select, organise and present information to people seeking answers about political issues.

AI chatbots becoming an important source of election information

The electoral chief noted that Canadians are increasingly using AI chatbots to obtain direct answers to questions about the voting process and political choices.

These systems can summarise information from multiple sources and respond to questions about registration, voting procedures and competing political platforms.

Although the technology offers potential benefits by making information more accessible, Elections Canada warned that the answers provided may not always be accurate, complete or based on authoritative information.

The Chief Electoral Officer expressed concern that users may not know which sources a chatbot relies on when generating its responses.

He also warned that AI systems could influence what voters understand about political parties and candidates without making the process of selecting and presenting information sufficiently transparent.

For example, a voter asking an AI chatbot about where to vote or how to register may receive an answer that is incomplete or inaccurate if the system relies on outdated or unreliable information.

The concern is that voters could make decisions based on information that appears authoritative but has not been properly verified.

Elections Canada calls for greater transparency

One of the recommendations presented to parliamentarians was the introduction of stronger transparency requirements for AI-generated political communications.

The Chief Electoral Officer said he had previously recommended that synthetic content used in electoral communications should carry clear labels identifying it as AI-generated or manipulated.

He indicated that this issue falls within the scope of Bill C-34, which is being considered in the context of transparency in electoral communications.

The electoral authority also wants AI platforms and chatbot providers to direct users to official or authoritative sources when answering questions about voting procedures.

This would help ensure that people seeking information about when, where and how to vote can verify the answers they receive through reliable channels.

The Chief Electoral Officer further recommended broader disclosure requirements for electoral communications issued by political parties, candidates and third-party groups.

He argued that existing requirements, which focus largely on election advertising, do not adequately cover the wider range of political communications circulating online.

Such communications can include videos, social media posts, digital advertisements and other forms of content that may influence voters without necessarily appearing as traditional campaign advertising.

Deepfakes remain a significant concern

Although the electoral chief believes the threat goes beyond deepfakes, digitally manipulated content remains a major concern for election authorities.

Artificial intelligence can be used to create convincing videos, photographs and audio recordings that make it appear as though political candidates or public officials have said or done things that never happened.

Such material can spread rapidly through social media, potentially misleading voters before election authorities or fact-checkers have an opportunity to respond.

Elections Canada has previously warned that AI-generated content could be used to impersonate candidates, election officials and other public figures.

The agency has also highlighted the possibility of foreign actors using the technology to spread disinformation, create confusion and undermine confidence in democratic institutions.

In its guidance on digital literacy, Elections Canada distinguishes between misinformation, which involves sharing inaccurate information without necessarily knowing it is false, and disinformation, which involves deliberately spreading false or misleading material.

The agency says knowingly publishing false or misleading information with the intention of disrupting an election or affecting its outcome can violate the Canada Elections Act.

Experience from Quebec raises further concerns

The latest warning follows research presented to Canadian parliamentarians in October 2026 about the use of AI chatbots during Quebec’s recent provincial election.

A study led by McGill University political scientist Aengus Bridgman estimated that approximately 40 per cent of Quebec voters had turned to AI chatbots for election-related information.

However, researchers found that some of the information provided by the systems was inaccurate or outdated.

Taylor Owen, an associate professor at McGill University’s Max Bell School of Public Policy, told a parliamentary committee that a significant portion of voting advice provided by ChatGPT during the final weeks of the Quebec campaign came from an error-filled website created using AI.

Researchers said the situation demonstrated how AI systems could unintentionally distribute misleading information when reliable news organisations and authoritative sources were not accessible to them.

The findings raised questions about how much confidence voters should place in automated systems when seeking information about candidates and political parties.

They also highlighted the need for AI companies to improve how their systems identify reliable information, particularly during election periods.

Elections Canada says it will not use AI to count ballots

While raising concerns about the risks associated with artificial intelligence, the Chief Electoral Officer also explained how Elections Canada is approaching the technology internally.

He said the agency was finalising an internal AI strategy focused on employee awareness, governance, engagement and cautious exploration of potential applications.

Importantly, Elections Canada has no plans to use artificial intelligence in the actual voting process, the counting of ballots or the provision of information directly to voters.

Instead, the agency is considering possible administrative applications, including summarising documents, assisting with software development and supporting certain geographical functions.

The electoral authority also intends to publish a transparency statement explaining its use of artificial intelligence.

The Chief Electoral Officer said maintaining the integrity of the electoral process and public confidence in elections would remain the agency’s overriding priorities.

Why the warning matters to Canadian voters

The growing use of artificial intelligence is changing how people access news, compare political positions and engage with public affairs.

For voters, the technology can make complicated information easier to understand and provide quick access to material that might otherwise take considerable time to research.

However, the risks increase when people accept AI-generated responses without checking the information or examining the sources behind them.

A chatbot may present an answer confidently even when the information is incomplete, outdated or inaccurate.

The concern raised by Elections Canada is therefore not simply whether voters can identify a fake video. It is whether they can understand how the information shaping their political opinions is selected, presented and distributed.

As Canada continues to examine the relationship between artificial intelligence and democracy, the recommendations presented to Parliament could influence future rules governing political communications, online platforms and the transparency of AI-generated content.

The challenge for policymakers will be to protect voters from manipulation while preserving legitimate political expression and access to information.