Morning Wire - Threat of Social Media Warfare | Sunday Extra

Episode Date: November 19, 2023

We speak to a threat intelligence expert about how fake social media accounts are being used to promote propaganda and influence vulnerable audiences. Get the facts first on Morning Wire. Learn more about your ad choices. Visit podcastchoices.com/adchoices

In the midst of Israel's conflict with Hamas and Russia's invasion of Ukraine, the importance of discerning truth from lies on social media platforms has become critically important. TikTok in particular is where misinformation appears to be targeting younger and more vulnerable audiences. In this episode, we speak to an expert on identifying online influence operations and combating fake social media accounts. I'm DailyWire, editor-in-chief John Bickley, with Georgia Howl. It's Sunday, November 19th, and this is a day. an extra edition of Morning Wire. Joining us now to discuss the use of fake accounts

surrounding recent events is Rafi Mendelsohn Vice President at Saiabra. Rafi, welcome on. So first, what does your company Saiaabra do? So we are a social threat intelligence platform. What that means is that we work with governments and also private companies to trap conversations on social media,

all publicly available information, and to be able to detect disinformation and threats across social media by analyzing the behavior of profiles and especially identifying fake accounts, bots that exist across social media. And so we're able to identify fake accounts and then track them and see the impact that they are having on the wider conversation as well. How difficult is that to do to figure out whether or not something is a fake account? I know sometimes things are flagged as bots that actually aren't. Is this something you guys feel very confident about when you flag accounts?

It's incredibly difficult, which is why no one else has really done it. The co-founders of the company come from the information warfare units of the Israeli Defence Force. And so this is the area that they worked in. And then five years ago when they created Saiba, they specifically set out to create a platform that is designed to analyze and identify the behaviors of malicious actors and so that we can see what's going on. Fake news and disinformation is quite a general term. But what we are often focused on specifically are the fake accounts and the influence operations

efforts that are being conducted by malicious actors. So, for example, we've worked with the US State Department to identify foreign interference with elections. We work with large companies to track stock price manipulation or impersonations of companies or senior executives. And so we feel our platform uses machine learning algorithms. to be able to detect the behaviours of accounts. So we don't look at content, right,

where we analyze content for sentiment, but we don't say, oh, John has posted that the sky is red and actually he's not correct. What we're looking at is hundreds, between 6 to 800 different behavioral parameters to say, you know, is John behaving like an authentic person online?

Are you posting 23 of the 24 hours of the day? Oh, okay, that's something that might, flag up in our algorithms that allows us to say with confidence that this is an inauthentic profile. That could be a sock puppet or a bot or a troll, etc. But those are the accounts and the kinds of profiles that we believe that we should be monitoring really carefully because most of the time they've been created with the purpose of nefarious intent. Since October 7th, the Hamas attacks, what types of inauthentic behavior are you seeing and where is it coming from? Wow. So since the seventh, as you can imagine as a company, we've tracked times of war. We were very closely monitoring Russian bot networks when they invaded in Ukraine. We monitor the US 2024 elections. We're finding a lot of fake accounts. This is what we do as a company. So when October the seventh happened, we started focusing our technology on this topic of conversation. And what we saw in terms of the scale of the fake accounts,

the sophistication of the influence campaigns that were being run, the planning that went into this, some of these fake accounts that we have uncovered were created a year, a year and a half before the conflict, lying pretty much dormant, not really doing anything. And then on the 7th of October, they were turned on, as it were. Are you finding that a lot of states, a lot of governments are involved in this bot business? Well, whether it's a lot of states or one state or the scale of the organizational support. We've uncovered over 40,000 pro-Hamas coordinated fake accounts across the main social media platforms.

But almost immediately, from the 7th of October, we're pumping out false narratives in various languages for various audiences. In the first two days, on the 7th and 8th of October, we identified that 1 in 4, 25% of the conversation around the conflict had fake accounts. that was driven by fake accounts. And so as any war starts or as any global event starts nowadays, we go towards social media, right?

We go to social media to see the news, to see the breaking news. We do as a population generally, journalists do politicians do themselves. And one in four accounts involved in this conversation in the first two days was a fake account. And so we should be really kind of suspicious of that, but also we should be alert as to the scale of what's being conducted by the other side. Yeah, I think a lot of people will be very shocked to hear that statistic.

Is there a particular platform or platforms where this is particularly bad, rampant? Well, we see it across all the main platforms, and I think the best way to think about it is that if you're a malicious actor, you're going to go where the eyeballs are. TikTok because of the nature of video content and the way that you engage around different types of content and it tracks how long you spend, how long you're engaging in one video or another,

and then it serves you more of the videos that it thinks that you like. And with this kind of conflict, even from the first day, the amount of horrifying videos that were being shared around, that is prone to abuse on X on Twitter. We see there are lots of fake accounts rife there. But equally on Facebook and Instagram, we also see a really troubling activity there as well. So it's across all the main platforms,

even though there might be slightly different. Do you feel like you've made progress in terms of weeding out some of this. Are things getting better or worse on this issue? I wish that we could say that it's getting better, but unfortunately, I think we are still behind as a society, politicians, of being able to understand the true scare of what's happening here. There is a conversation to be had about, you know, free speech on social media, what people can say, how they say it's, you know, that's a conversation. But at Saiba, what we are particularly focused on, is at least understanding who we are speaking to is actually a real person and not a bot or a troll or a fake account.

In terms of individuals and how they operate online, do you have advice on how to recognize inauthentic behavior? Absolutely. I think for the average consumer social media, I think the advice would definitely be to take another moment, understand what you're saying, especially when some of these images and videos that we're seeing are so emotive and takes our breath away some of them in such a negative way, it's easy to be swept up by that. But taking a moment to really look at what we're seeing, but also as well, not just the content, but the profile itself. And then maybe take that time to go onto the profile. Was this account created a week ago?

Is it posting every few minutes? Is the behaviour that it's showing authentic or inauthentic? That would be, I suppose, the advice for people just to take an extra moment and be sensitive and critical. Social media can be a great place to join in specific conversations and engage with people. But taking that extra moment, knowing what we know, now with events like these with wars with conflicts, there can be a very high proportion of the conversation that is filled and driven by fake accounts.

Final question. Specifically about the war in Israel, what are some of the false narratives you saw coming out after the Hamas attack? One of them in the very early days was a false narrative about the compassion and humanity that was being shown by the terrorists towards the hostages, and that was being pumped out in English. But another false narrative that we saw being pushed out by thousands of fake accounts was the accusation that the imagery of Israeli deaths were actually created using AI. And that Israel had utilised AI generated images following the Hamas attack. And actually one of the accounts that we saw are being attacked more than most is Ben Shapiro's post, actually, from a couple of weeks ago, where on October the 14th, he posted an image of, of what all proof, as he put out there, of unfortunately,

was a baby that had been killed and burned. And we saw a lot of the focus from those fake accounts being directed towards him. Instead of spending millions of dollars on tanks and airplanes, they can actually spend a fraction of that price to be able to create programs and hire people to be able to create those false narratives and run those very, very sophisticated campaigns.

And it's so effective. It's working. It's reaching 100. hundreds of millions of profiles every week. And it's very difficult to detect and shut down unless we as a society come together and try and find the answers to this. A very complex new battlefront for sure. Rafi, thank you so much for talking with us. That was Saiabra Vice President Rafi Mendelssohn, and this has been an extra edition of Morning Wire.