Campaign May 29, 2026
The @HateTags campaign for The Toronto Holocaust Museum used the tricks of hate merchants’ nasty trade against them.
by David Brown | May 29, 2026
On May 12, new unbranded “ads” started appearing around hateful digital content on social media. The @HateTags ads weren’t for a product or service, but were instead designed to get people about to watch the content to think critically and “view with caution,” because hate and misinformation that starts online can cause real world harm.
The agency behind the campaign, Diamond, used some of the tools used by people to spread hate content—bot farms, keyword targeting and comment infiltration—to post the warning labels on the content, much like health warnings on cigarettes and other addictive products. Diamond also took on hate merchants by using other tricks of their nasty trade: posting in the comments, tagging content on Instagram and TikTok, and using bot farms to boost comments that supported the cause and increased visibility.
The campaign for The Toronto Holocaust Museum comes in response to the well-documented proliferation of hateful and toxic content and misinformation through ad-supported digital media. Along with pushing viewers to think differently about what they are being fed by their algorithms, it also calls on the Canadian government to catch up with other governments around the world and require the platforms to do more about spreading such toxic content.
“As an institution educating on the enduring lessons of the Holocaust, we deeply understand that when hate, propaganda, and misinformation spread, the consequences can be devastating,” said the museum’s executive director, Dara Solomon. “Its legacy serves as a broader warning about the real-world dangers of unchecked hate and the responsibility we all share to confront it.”
In total, 11 warning labels were placed across YouTube, Instagram and TikTok, with non-skippable labels on YouTube viewed more than 480,000 times. The campaign ran across Canada, but was shared beyond borders. The campaign’s landing page, HateTags.com, has received visitors from 30 different countries.
The campaign, which did not mention the Museum at first, has moved to the next stage with the introduction of a petition:
“The Museum is now urging lawmakers to mandate these warning labels across all social media platforms. Sign the petition to drive the federal government to introduce legislation that mandates:
- All social media platforms implement warning labels on content identified as hate speech or misinformation.
- Establish national guidelines and definitions of hate speech and misinformation.
- Ensure that warning labels are prominent, accessible and provide clear warning to help users think critically about the content they are consuming.”
Campaign wrote a short news story when the campaign was revealed earlier this week, but asked the Diamond CDs behind the campaign, Jordan Cohen and Mark Holden, to tell us more about how they pulled it off.
Was there a starting brief from the museum that led to this, and how did you arrive at this media-led campaign instead of a typical ad campaign?
The Toronto Holocaust Museum received a grant from Heritage Canada for a national campaign to combat online misinformation that fosters hate. The job was to get young Canadians to think critically about news and content served up to them.
We needed to show them how easy it is to be manipulated by the algorithm by getting sucked into conspiracy theories. That was going to take an intervention, not a PSA.
As we kept digging into the problem, the conversation kept coming back to real impact. Just talking about the problem wasn’t enough — we wanted to actually go where the problem lives, which is inside the algorithm itself.
So the idea became about using media spend to disrupt that cycle in real time. Getting in front of someone right as their feed starts pulling them toward misinformation. Creating an actual interruption, not just another piece of content to scroll past.
How did Diamond arrive at this specific idea?
It started with a pretty simple observation: the way people consume misinformation online looks a lot like how people used to smoke. It’s automatic. Mindless. You’re not really choosing to do it—you just kind of are.
Cigarette warning labels cracked something important. They work because they force a connection between the action and the consequence, right at the moment it’s happening. We wanted to leverage that logic and bring it online.
That’s what HateTags are—warning labels for misinformation. Something that makes you pause for a second and actually reckon with what you’re watching and sharing before you do it. Most people have had the thought that social media should have warning labels. We just actually did it.
As we said this feels like a very media-led idea, what can you tell us about Epitaph’s role here?
Epitaph was a crucial partner in making the technical side of this work. They brought in Silverpush, a brand safety AI that advertisers normally use to avoid running ads next to objectionable content. The whole point is usually avoidance, but we flipped that completely. Instead of running from hate content, we used the technology to find it—and run directly before it. That’s what made the intervention possible at the exact moment it needed to happen.
@Hatetags had no logo, tagline, or obvious museum branding. Why was that?
This was a deliberate strategic call. And one that took real courage from the Toronto Holocaust Museum.
We wanted Hatetags to be able to combat all hate. Racism, Islamophobia, antisemitism, the manosphere—all of it.
If we launched with the Museum’s logo attached, the conversation immediately narrows. People start asking “why is a Jewish institution talking about this?” instead of engaging with the message itself. So we launched anonymously under Hatetags. No institutional branding. The tags needed to feel like they belonged to everyone.
Here’s what makes that remarkable: the Toronto Holocaust Museum funded this, they championed it, and they were willing to step back from the credit because the mission mattered more than the marquee. That’s not a small thing—especially right now, when that community is absorbing hate constantly.
Now that it’s publicly launched, they’re fully, visibly connected to it. As they should be.
How exactly did the AI identify hate content?
We used Silverpush’s AI-powered video scraping technology—the same tool brands use to protect themselves from appearing next to objectionable content—and reverse engineered it for the opposite purpose. Instead of avoiding hate, we were hunting for it.
The ads were contextually targeted based on visuals, audio, copy, comments, and hashtags across a range of hate and misinformation categories.
Working closely with the Toronto Holocaust Museum, we built a framework of keywords, themes, and content patterns tied to the specific types of hate the campaign was designed to address. The detection is automated, but the outputs are reviewed and refined on an ongoing basis to keep the targeting sharp.
What was striking throughout the process is how easy it was to find this content—even without any special access to the platforms themselves. Which raises an uncomfortable question: if we could do this with a media buy, what could the platforms do with all the data they’re actually sitting on? They have infinitely more access than we did. They choose not to use it the same way.
What were the greatest challenges?
The biggest challenge was figuring out how to operate inside spaces that are already being weaponized by people spreading hate.
Early on, we made a decision: if bad actors are using aggressive tactics to manipulate these platforms, we need to understand those same systems well enough to interrupt them. Use their tools against them.
That meant building and deploying bot farms, navigating shadowbans, and staying ahead of platform detection systems. Platforms are specifically designed to suppress exactly what we were doing—so it required patience, rigorous testing, and genuinely creative technical thinking to keep it working.
Was it against platform rules? Yes. But so is coordinated hate, and that’s still online. We made a conscious decision that the cause justified it—and we’d make it again.
