“Taylor Swift is a Pentagon asset.” How do such crazy narratives get so popular?

“Taylor Swift is a government asset to shift people's voting preferences” is just one of the crazy narratives being spread online, and some actually believe it. How does something like this happen? As we attend the Black Hat USA 2024, we try to find out more.

States and private organizations alike go to crazy lengths to attack each other. Did you know that a certain company might pay someone $50,000 to organize an informational attack against their competitor?

Or that such crazy fake news like “Zelenska bought a Bugatti hypercar” or conspiracies like “Taylor Swift is a government asset put there to influence young voters” spread like wildfire not because people actually find it believable but because there’s someone behind it?

With AI’s help, it’s become much easier to spread false and misleading narratives. Operators behind conspiracies no longer have to try to convince people of something, instead, they deploy bots to exaggerate a certain narrative and influence people and decisions.

Experts at Blackbird.AI, a security company specializing in brand reputation protection, use the term thoughtnet, saying it’s an emerging threat that combines targeted cyberattacks, AI content, and bots to undermine security and trust online.

This week, at the Black Hat USA 2024 conferences in Las Vegas, we met with the company’s CEO, Wasim Khaled, to learn more about thoughtnet and the curious engineering behind the rapid spread of narratives.

Blackbird.AI specifically focuses on protecting organizations from narrative attacks that can essentially force firms into making multibillion-dollar decisions based on fabricated discourse or be otherwise hurt by amplified false narratives.

This interview has been edited for length and clarity.

Could you elaborate on how thoughtnet is different from other more traditional forms of propaganda?

Every organization today has a narrative. And then there is a whole set of counter-narratives that can spin up to create confusion and manipulate viewpoints in the public and even within areas of government. Anybody can be subject to a narrative attack – a person, a place, a company, a product. An attack is an assertion in the information ecosystem that can drive massive reputational or financial harm to any number of entities.

What would a successful narrative attack look like?

When we go to the private sector, which is where most of our work today lies, it's about understanding the public's perception of a company, an organization, or a product. Everything from their stock price and market manipulation to the things that a CEO might say can then be taken and subverted into something that can be controversial or drive fabricated controversy.

I'll give you a few examples in the enterprise space. There was a consumer product goods company, a really large global one that, during the beginning of the Russia-Ukraine war, had to make a decision on whether or not they were to stay in the country or leave. And they provided a product that actually was really important to children and infants. They didn't really want to just leave because then there would be a narrative that they did harm. But there was a massive outcry to have them leave, much like those other thousand companies that were facing a boycott.

In this case, we identified that most of the conversation was bot-driven. Bot-driven is an amplified narrative that wouldn't have gotten the reach or attention, but it was fabricated by people who found a gap to drive an agenda. In this case, they wanted this company to leave Russia, and they found that all of that discussion was fabricated. So they actually decided to stay and keep providing those products. That was a multi-billion dollar decision that they made virtually overnight.

Across almost any sector, you have to know the narrative. You have to understand the counter-narratives. You have to understand the contagion-like effect in which narratives spread, essentially like a virus infecting anyone it touches. You have to understand if there are deepfakes or generative AI.

If the narrative is changing, then you have to understand bot networks or synthetic amplification of that narrative. Then, think about the actor groups that exist. So you're going to want to know if state actors, for example, are amplifying or present and kind of converging with other groups you might not want, paying attention to your organization.

Do you focus on who's behind those attacks? If there are groups trying to influence the narrative, and if so, what are those groups?

We use ‘cohorts’. These are kind of tribal communities that are like minded and talking about the same things, driving the same kinds of narratives. So first of all, in terms of attribution, if it's like state actors, let's say it's Russian state actors, CCP (the Chinese Communist Party) state actors, we would identify these groups of accounts that are driving those narratives so they understand that's the intent.

When we think about attribution, it's really not like this is the individual sitting in a particular place in the world – it's more about the narratives and what types of actors are driving them.

If it's like someone in public health, we might say there's an anti-vax cohort that is driving this narrative. Right. With banks, we've seen, say, anti-capitalist versus capitalist narratives.

Most of our customers, that's really what they're getting at anyway. It's like, what are the hidden mechanisms that we didn't see that create risk for us? [...] There's no magic button to unwind these kinds of nuanced, harmful narratives. It's about knowing exactly what's happening. So you can respond accordingly.

How do botnets tie into the whole thing?

People use the bot attack because otherwise, you have to convince someone this is a problem and have them create a conversation around it. Here, you can take any kind of opportunity, which is like any kind of incident. And then it could be a real person, it could be true, it could be false. You can amplify it so that it gets a lot of attention and a lot of reach.

There was this scenario with Taylor Swift in the US before the Super Bowl.

One in five Americans believes that Taylor Swift is a government asset that was put there to shift people's voting preferences. How does something like that happen?

Right before the Super Bowl started, the fringe web said this was the case – she's a Pentagon asset. They recontextualized some things that they had seen in the news, whoever was on the dark web there, and this conspiracy just caught on fire. Some of those narratives speculated on the relationship between Travis Kelce, the football player, and Taylor Swift, saying they were created by the government to influence young voters.

So now you have this kind of football fan that that cohort around the NFL, and you have her fan base, which is massive. These caught on. And because no one was really able to understand how that spread, suddenly you just had millions of people getting exposed to this, and this is just what they believe. Now, all of this had, you know, your NFL fans, your Taylor Swift fans, but it also had lots of Russian state actors amplifying those conspiratorial narratives to reduce trust in the government.

There was a similar scenario, equally outlandish, where an alien was discovered in Mexico. I got a lot of calls about that one, too. That was essentially also amplified as here's another thing that the government has hidden from you.

I give those examples because, regardless of what it is, it erodes trust in the government. Then you have narratives like US taxpayer dollars are going to fund Zelensky's yacht. This is a bigger one because it involves elections, funding for war, and all of those things.

We were at the White House briefing the National Security Council in December on these Zelensky narratives. There's another one around trying to convince Americans that they're going to be drafted into the Ukraine war.

There are just a lot of things that you can do with this tradecraft that can be applied to the public sector, private sector, and society at large. Without having the technology to fight, detect, and mitigate against these technology-driven attacks, it's like not having anti-ballistic missiles for nuclear warheads. You're just wide open, right? So you're fighting blind. And that's really what we provide. We provide this kind of radar for this kind of thing and ways to address it.