Published 11:14 IST, May 25th 2019
Fake Facebook accounts: The never-ending battle against bots
The staggering figure of more than three billion fake accounts blocked by Facebook over a six-month period highlights the challenges faced by social networks in curbing automated accounts, or bots, and other nefarious efforts to manipulate the platforms.
Advertisement
staggering figure of more than three billion fake accounts blocked by Facebook over a six-month period highlights challenges faced by social networks in curbing automated accounts, or bots, and or nefarious efforts to manipulate platforms.
Facebook said this week it "disabled" 1.2 billion fake accounts in last three months of 2018 and 2.19 billion in first quarter of 2019.
Advertisement
Most fake social media accounts are "bots," created by automated programs to post certain kinds of information -- a violation of Facebook's terms of service and part of an effort to manipulate social conversations. Sophisticated actors can create millions of accounts using same program.
Facebook said its artificial intelligence detects most of se efforts and disables accounts before y can post on platform. Still, it ackwledges that around five percent of more than two billion active Facebook accounts are probably fake.
Fake accounts may be used to amplify popularity or dislike of a person or movement, thus distorting users' views of true public sentiment.
Advertisement
Bots played a disproportionate role in spreing misinformation on social media ahe of 2016 US election, according to researchers. Malicious actors have been using se kinds of fake accounts to sow distrust and social division in many parts of world, in some cases fomenting violence against groups or individuals.
Bots "don't just manipulate conversation, y build groups and bridge groups," said Carnegie Mellon University computer scientist Kathleen Carley, who has researched social media bots.
"y can make people in one group believe y think same thing as people in ar group, and in doing so y build echo chambers." Facebook says its artificial intelligence tools can identify and block fake accounts as y are being created -- and thus before y can post misinformation.
Advertisement
"se systems use a combination of signals such as patterns of using suspicious email dresses, suspicious actions, or or signals previously associated with or fake accounts we've removed," said Facebook analytics vice president Alex Schultz in a blog post.
figures from Facebook's transparency report suggests Facebook is acting aggressively on fake accounts, said Onur Varol, a postdoctoral researcher at Center for Complex Network Research at rastern University.
Advertisement
"Three billion is a big number -- it shows y don't want to miss any fake accounts. But y are willing to take a risk" of disabling some legitimate accounts, Varol said.
Legitimate users may be inconvenienced, but can generally get ir accounts reinstated, researcher ted. "My feeling is that Facebook is making serious efforts" to combat fake accounts, he ded.
But new bots are becoming more sophisticated and harder to detect, because y can use langu nearly as well as humans, according to Carley.
Advertisement
"Facebook may have solved yesterday's battle but nature of se things is changing so rapidly y may t be getting new ones," she said.
Varol agreed, ting that "re are bots that understand natural langu and can respond to people, and that's why it's important to keep research going."
Many users can't tell difference between a real and fake account, researchers say. Facebook and Twitter have been stepping up efforts to identify and weed out bogus accounts, and some public tools like Botometer developed by Varol and or researchers can help determine likelihood of fake Twitter accounts and followers.
"If you use Facebook to communicate with family and friends you should t worry much," said Filippo Menczer, a computer scientist who researches social media at Indiana University.
"If you use it to access news and share that with friends, you should be careful." Menczer said many Facebook users pay little attention to source of material and may end up sharing false or misleing information.
"Everyone thinks y cant be manipulated but we are all vulnerable," he said.
Along with bots, humans represent a key element in misinformation chain, researchers say. "Most false information is t coming from bots," Carley said. "Most of it comes from blogs and bots rebrocast it" to amplify misinformation.
Facebook chief Mark Zuckerberg said Facebook is seeking to eliminate financial incentives of fake accounts.
"A lot of harmful content we see, including misinformation, are in fact commercially motivated," Zuckerberg told reporters. "So one of best tactics is removing incentives to create fake accounts upstream, which limits content me downstream."
11:14 IST, May 25th 2019