Information Incident Notification: Kirkland Lake Bot Campaign
Overview
On July 31st, Conservative Leader Pierre Poilièvre's rally in Kirkland Lake attracted a large audience as part of his Northern Ontario tour.
On August 3rd, online engagement focused on the rally surged with hundreds of posts on X (formerly Twitter) from over a hundred accounts. Some individuals and media have suggested these were bots.
This quickly raised widespread speculation about the origin of these accounts and whether a specific political party or third party actor was responsible.
Given that the accounts bios were primarily based from outside Canada, there was also speculation of whether their posts constituted foreign interference.
If this bot-generated campaign was initiated by any federal party, it raises ethical concerns about whether they are attempting to influence Canadians. Incidents like these can mislead the public and affect trust in democratic processes.
On August 9, 2024, the CDMRN research team activated their information incident research response team to conduct investigation on the nature of the information incident, characterize the impacts, and identify lessons learned.
Over the coming weeks, the CDMRN will release incident updates sharing findings of our research on topics including diagnostics, impact and foresighting.
Timeline of the event
July 31, 2024
Conservative Leader Pierre Poilièvre holds an Axe the Tax Rally at the Heritage North Conference and Events Centre in Kirkland Lake, Ontario, a town of 7,750 people.
August 1-3, 2024
The event receives local news coverage, shared by Poilievre on X, including news items by CTV Northern Ontario, TB News Watch and in the Kirkland Lake Northern News.
Screenshot by X user The280Times Trivia Channel (@The280Times)August 3, 2024
In the late afternoon, accounts on the X platform start posting messages claiming to be from Northern Ontario and to have attended the rally. The posts feature highly similar wording (sometimes identical), many referring to having “just [returned/got back]” and to feeling “refreshed” after the event.
Users–particularly @The280Times–quickly identify the sudden surge of “refreshing” Kirkland Lake messages and find several hundred posts from over a hundred accounts, most of which seem to belong to recently created accounts from users residing outside of Canada, despite language in most messages that identify them as “Northerners” or “Northern Ontarian”.
August 4, 2024
NDP MP Charlie Angus shares a post on X by Luke LeBrun of news outlet Press Progress about the flood of claims about attending the Kirkland Lake event.
August 6, 2024
The NDP issues a press release that questions whether the Conservatives used bot farms and asks the Conservatives to explain why these “offshore bot farms are working to influence Canadian voters on Poilièvre’s behalf”.
August 7, 2024
NDP requests that Canada’s Election Commissioner investigate whether online bots were used to back Pierre Poilièvre as “bot farms have been identified as part of broader campaigns to undermine democracies.” The Office of the Election Commissioner confirms receiving the letter but would not say what steps the office would take, if any.
A group of Liberal MPs from Northern Ontario write an open letter to Elon Musk to object to “blatant online manipulation designed to influence public opinion with disinformation and to interfere in Canada’s public discourse and democratic processes.” They ask that the company share evidence and information about the accounts and the campaign.
The Conservative Party of Canada (CPC) denies any connection to the posts: according to a representative, “The CPC does not pay for bots and has no idea who is behind these accounts.”
Key questions we are focusing on
Many questions emerge that merit closer investigation. These include:
How aware are people of the event?
How many people are aware?
What do people believe about the event?
How did they find out about it?
What do people believe about the event?
What do they believe happened?
Where do people think this came from?
Are people concerned about incidents like these?
How significant was this event?
Where lies the largest impact? How did the Kirkland event compare to others? If bots were used, what is the impact of the bots versus the indirect impacts generated by discussion about the bots?
How does this event compare to a worst-case (or worser-case scenario)?
What would a more severe-case scenario look like?
What can we learn from this event that can inform future planning and response?
Research response
The CDMRN is conducting research to understand the situation in more depth, and characterize its impacts on the Canadian digital ecosystem. This study will use a combination of techniques including survey, social media analysis, and generative AI analysis. In the upcoming weeks, we will release incident updates on topics including:
Diagnostics of this information incident - insights on public awareness of the event (e.g., did people know about event, how did they find out about it) and on beliefs (e.g., what do people think happened, where do they believe it originated from, who they think is responsible), as well as more insight on the use of bots in politics online.
Impacts of this information incident - insights into how it has affected public perception (e.g., should Canadians be concerned, does it impact their trust in politicians, media, or the electoral process), and what factors generated the most impact on Canadians perception (e.g., was it the bot-generated campaign or the media coverage about it).
Foresighting information incidents - insights on how this incident could compare to a larger, more coordinated bot campaign leveraging generative AI techniques.
Other updates might be provided as the event unfolds.
Research partners
This incident response notification is a CDMRN collaboration between the Applied AI Institute, the Network Dynamics Lab, the PolCommTech Research Lab, the Media Ecosystem Observatory and the Social Media Lab.
Key contacts - Subject matter experts
Alexei Abrahams, Digital Lead, Media Ecosystem Observatory (alexei.abrahams@mcgill.ca)
Expertise: Social network analysis/Digital trace, media observatoriesAengus Bridgman, Director, Media Ecosystem Observatory (aengus.bridgman@mcgill.ca)
Expertise: Political science, social network analysis
Derek Ruths, Data scientist, Network Dynamics Lab (derek@derekruths.com)
Expertise: Social network analysis, social media analysis, hate speech, online user behavior studies
Elizabeth Dubois, Director, PolCommTech Research Lab (elizabeth.dubois@uottawa.ca)
Expertise: Bots
Fenwick Mckelvey, co-Director, Applied AI Institute (fenwick.mckelvey@concordia.ca)
Expertise: Generative AI
Jennie Phillips, Project Director, Project on Infrastructure Ecosystem Resilience, Media Ecosystem Observatory, (jennie.phillips@mcgill.ca)
Expertise : Research Project lead/research design
Philip Mai, Senior Researcher and Co-Director, Social Media Lab, (philip.mai@torontomu.ca)
Expertise : Social network analysis