A new survey from data security firm Imperva reports that it is non-human for half of all internet traffic in 2024, with 37% being bad bots, up 5% since 2023.
Minneapolis – Last July, an assassin attempted to seize President Donald Trump’s life at a Pansylvania rally.
Within 48 hours, the conspiracy theory was ramped on social media. One of them suggested that Trump staged the case to boost his support for the election.
Cyabra, an artificial intelligence disinformation detection company, has analyzed thousands of profiles related to the story. They claim that 45% of them are fake accounts aimed at spreading misinformation, reaching a potential 595 million people.
“Within minutes of that attempt, we saw many fake accounts that were driving a specific, completely false story, but they managed to engage in the conversation,” said Rafi Mendelsohn, Cyabra’s chief marketing officer.
Cyabra uses AI to catch bots by analyzing the behavior and content of social media accounts in real time to determine whether they are real or fake. Mendelsohn says that he uses behavioral parameters between 600 and 800 when analyzing an account.
Vowing to “try to defeat spambots and die,” Elon Musk hired Saibura in 2021 and investigated how bad the issue was on Twitter before purchasing the platform.
Cyabra concluded that 11% of Twitter users are spam or bot accounts.
Now that X is under mask ownership, Cyabra’s research claims that one in five accounts that interacted with Musk’s 2024 election-related post is fake.
Multiple reports show that BOT activity on social media is generally more refined.
“The adoption of AI tools by malicious actors has made creating fake accounts much more effective and efficient,” Mendelsohn said. “It’s really kind of a mall situation.”
Bots today come in all forms and intents.
There are so-called “good” bots like AI customer service agents and marketing bots, supporting people and businesses with tasks.
And then there are “bad” bots designed for malicious attacks on your identity, your money, or your way of thinking.
A new survey from data security firm Imperva shows that over half of all internet traffic in 2024 was human, with 37% being bad bots, up 5% since 2023.
“They are changing the structure of society,” said Hagai Shechter, CEO of Furaudlogix. “They are changing the way we think about certain things, how we think about certain things.”
Furaudlogix monitors internet activity to alert advertisers if clicking on an ad or site is from a real person.
Schector says that bad bots have matured, but now they can think, learn and chat for themselves.
“Imagine you’re in a room with 1,000 people and someone says something and the whole crowd starts to applause. That kind of thing affects how you feel about it.
Kare 11 Test
Have you ever scrolled on social media and thought, “It has to become a bot”?
Recently I’ve been wondering what the bot looks like and how many of these fake accounts are out there.
I asked Cyabra to use the software to analyze several recent Kare 11 posts in hundreds of comments.
This is in a post about X on the deportation of immigrants convicted of fraud, and Cyabra’s analysis found 225 fraudulent accounts. This means you’re more confident that 36% of all comments about your post are bots.
Their report allegedly expressed negative feelings towards deported criminals or positive feelings towards ice and law enforcement.
In the second post about targets being drawn from the Twin City Pride Parade, Cyabra’s report identified 32% of comment accounts as fake.
They provided some examples of suspected bot accounts to comment on the post.
Accounts with handles such as “Anonwhitedude2”, “robocop198011”, and “whynot28419645” have no photos, no names in their profiles, and few followers. Anonwhitedude2 had 16 followers. The account was created in January 2024 and frequently reposted racist, anti-Semitism, pro-Christian and pro-Trump content. Although there were no original posts on his profile, he often responded to political posts and also responded to posts from local sports and ’80s movies.
We reached out to all eight accounts that Cyabra suggested to be a bot. My Anonwhitedude2 account was deleted soon.
However, not all experts agree that these are bots.
Professor Darren Linville, co-director of Clemson University’s Media Forensic Hub, said:
He was also hired by Elon Musk’s legal team in 2021 to investigate Twitter bots.
Linvill believes that Cyabra’s bot estimates are inflated.
“It’s absolutely in their interest to exaggerate the problem. They gave you numbers. They gave you no real support for that number. They said these accounts were fake, but at the end of the DA you can’t see anything you can’t see,” Linville said.
In response to the claim, Cyabra said it supports that number.
“With the democratic intelligence reporting teams around the world, governments have been moving forward for a very long time to test the accuracy of their services and detection capabilities,” Mendelsohn said.
Both experts agree that the problems are on the rise.
Who is behind the bot?
According to Linvill, social media has three basic types of bot activities: marketers, scammers and political influencers.
“The largest number of bots are created by different types of marketing companies that create a large number of accounts that exist to promote other accounts,” Linvill said.
His research identified a network of bots during the 2024 election, creating more than 130,000 posts, targeting five US Senate races, one US house race and a presidential election.
“It’s not illegal. It’s against the platform policy. The simple act of creating a fake account and engaging in political propaganda on either the left or right is completely legal,” Limville said, unless the person creating the propaganda is a member of a foreign state.
We asked Linvill if leaders like X, Facebook, Instagram and Tiktok could stop all bots from being created on the platform.
He said they wouldn’t be able to do it and wouldn’t want to do that.
“One reason it’s difficult is because there are a lot of bots that exist legally. Perhaps NBC runs multiple bots to put out content. In fact, I know that. The platform creates policies that promote the presence of more accounts to make the platform more popular,” says Linvill.
For the big question, how many bots are there on social media? Linvill said there are no reliable estimates today as the account’s metadata is not public.
I contacted Meta and X to comment on this report. There was no response.
