“It’s Not an Anomaly” – Suspicious Reactions on a Post by the Minister of Immigration

The most recent Facebook post by Quebec’s Immigration Minister, Jean‑François Roberge, generated an unusually high volume of interactions: several thousand reactions, a significant portion of which came from accounts located in the Philippines, Pakistan, Indonesia, and other countries with no obvious connection to Quebec.

Many of these profiles also appear questionable — generic photos, no activity, or newly created accounts — raising concerns about the possible existence of fake accounts and bots, as well as the use of artificial amplification tactics.

This wave of engagement comes at a time when the reform proposed by the minister, transitioning from the Québec Experience Program (PEQ) to the Skilled Worker Selection Program (PSTQ), has faced strong criticism since its announcement. The reform is heavily contested by several concerned groups and poorly received by part of the Quebec population.

In this context, the unusual surge of positive reactions generated by a recent post on the minister’s Facebook account — totaling over 10,000 interactions — surprised many internet users, especially since Jean‑François Roberge’s posts usually generate only a few dozen reactions, and sometimes a few hundred at best.

La Converse analyzed a sample of 360 Facebook accounts that reacted to this post. The results raise serious questions.

According to an internal analysis conducted by Belinda Kanga, digital strategist and data analyst at La Converse, the positive reactions — “Like,” “Love,” and “Heart Man” — are linked to engagement behaviors that raise serious doubts. They also show significantly higher average suspicion scores than those observed in negative reactions. In contrast, accounts that reacted with an angry emoji appear largely organic. They come from real, consistent, and locally rooted profiles.

Analysis limitations

‍The investigation, conducted manually using an investigative approach, does not allow us to state with absolute certainty that these engagements were purchased or orchestrated. However, the observed trends indicate a probability level ranging from “high to very high.” Indicators such as foreign profiles, languages irrelevant to a Quebec context, semi-inactive or locked accounts, and inconsistent histories correspond to typical markers of “engagement farms,” which are linked to “bought engagement.” La Converse also consulted two independent computer scientists who examined these suspicious reactions and reached the same conclusions.

‍

“I can’t say with 100% certainty that these reactions are artificial, but the signals point to a high probability,” explains Belinda Kanga. “Two out of three reactions — and the most numerous ones — show characteristics typical of this type of activity.”

Two possible scenarios, according to an expert

‍This phenomenon is not entirely unprecedented in Quebec. According to Bruno Guglielminetti, a digital communications specialist and host of the podcast Mon Carnet, two scenarios are possible when a political post suddenly receives a large volume of suspicious reactions.

“Someone within the government may have made the poor choice of buying a few thousand likes, or someone in the opposition bought likes on this post to make the government look bad. This has unfortunately happened quite often,” he explains.

Guglielminetti also points out that a massive influx of reactions from fake or foreign accounts can “signal a distortion of public conversation.” According to him, this type of engagement misrepresents how the message is actually received and “muddies citizen opinion,” since the local audience is drowned out by external noise that could be manipulated.

He notes that this type of amplification can serve to “simulate popular support around a minister or a policy,” “influence general perception,” or result “from opportunistic networks of automated accounts exploiting highly visible content.” In some cases, it can also come from organized actors seeking to sway political discussion.

The expert warns of the risks linked to this practice: artificially inflated engagement can lead to “a loss of credibility, weaken trust in institutions, and damage the integrity of the public digital space.” He also notes it fuels generalized skepticism toward engagement signals, even when they are authentic.

Minister’s office denies paying for reactions

‍Contacted by La Converse, Minister Jean‑François Roberge’s office acknowledged noticing this “anomaly.” In an email to our newsroom, the minister also rejected the idea that he was behind the surge in reactions: “Alleging that we are responsible for the sudden increase in reactions is completely false.”

Furthermore, Mr. Roberge denies that any funds were spent on such an operation: “No money was used to buy visibility or reactions on this post. This is not how we operate,” his statement read. His office also noted that internal checks are underway “to understand the impact of the situation, including on the reach of our future posts.”

“It’s not an anomaly,” says the whistle-blower

‍For his part, Yacine Kraimia, the computer engineer who raised the alarm about this bot activity, considers it not an anomaly but a deliberate act. “I’m not saying the minister or someone in his office did it, but it’s not a mistake, it’s not an anomaly like the minister said — it’s something that was done, someone did it,” he stressed in an interview with La Converse.

Drawing on his expertise, Mr. Kraimia analyzed hundreds of reactions before noticing this suspicious spike. Calling the affair a “scandal,” he contacted the minister last Tuesday, then alerted the media about the use of bots or fake profiles on Minister Jean‑François Roberge’s Facebook page. “Using this type of tactic to artificially inflate public support is not only misleading but also deeply shameful for a government institution,” he wrote.

A challenge for digital transparency

‍This situation arises in the context of highly sensitive immigration debates in Quebec. The gap between the controversy surrounding the shift from the PEQ to the PSTQ and the surge of positive reactions observed on this post surprised many internet users. The high presence of foreign accounts makes the phenomenon even more striking.

As Bruno Guglielminetti also points out, there are “some tools” to detect inauthentic behavior, but no standardized public norms currently allow a clear confirmation or refutation of the true origin of such engagement. Platform mechanisms remain “opaque,” leaving these situations in a gray area that is hard to interpret. In a context of charged migration debates, this affair raises an open question: how can we ensure that political engagement signals still reflect something real? This is a fully-fledged democratic issue.

‍