:format(webp)/cdn.vox-cdn.com/uploads/chorus_image/image/54655079/675562678.0.jpg)
Friday night the campaign of leading French Presidential Candidate Emmanuel Macron reported it was the victim of a massive email hack just a day and a half before one of the most important elections in the history of France.
Macron is currently leading in the polls against his far-right rival Marine Le Pen.
The leaked documents have since spread like wildfire across social media, particularly on Twitter.
But, Nicole Perlroth, a cybersecurity reporter with the New York Times, pointed out that an overwhelming amount of the tweets shared about the Macron campaign hack appear to come from automated accounts, commonly referred to as bots.
5% of #MacronGate users account for 40% of tweets. The most prolific tweeted 1668X/24 hrs―faster than one RT per min, all day with no sleep.
— Nicole Perlroth (@nicoleperlroth) May 6, 2017
About 40 percent of the tweets using the hashtag #MacronGate, Perlroth noted, are actually coming from only 5 percent of accounts using the hashtag. One account tweeted 1,668 times in 24 hours, which is more than one tweet per minute with no sleep.
That account, along with the others in that 5 percent responsible for #MacronGate, appears to be automated, meaning there is no human at the helm but rather software.
Twitter appears to have been caught off guard with what is obviously a bot attack, despite the fact the social media company is well aware of the problem of bot accounts being used to falsely popularize political issues during high-profile campaigns to give the impression of a groundswell of grassroots support.
During the 2016 U.S. presidential election, bots sharing pro-Trump-related content outnumbered pro-Clinton bots by 7 to 1 during the third debate between the two candidates, according to research from Oxford University’s Project on Computational Propaganda. In the timespan between the first and second presidential debates, more than a third of pro-Trump tweets came from bot accounts.
Others have created third-party tools to determine if a Twitter account is a bot, like Bot or Not, a project from Indiana University that automatically scans for clues to determine if an account is automated. It looks at things like who the account follows and how it interacts with other accounts.
When Recode reached out to Twitter to ask what it is doing to combat the proliferation of bot generated tweets in light of the Macron campaign hack, a spokesperson pointed to policies the company has that prohibit the posting of automated tweets for trending topics. Twitter also has rules against the creation of multiple accounts to share redundant information. But those policies don’t seem to have stopped the outpour of bot activity underway following the Macron campaign hack.
It’s not clear what parts of the document dump are authentic and what parts aren’t. Macron’s campaign claims that fake information has been mixed in with actually hacked documents “in order to sow doubt and misinformation.” An investigation could take weeks.
There is evidence that Russia was involved with the hack. Last month, the cybersecurity firm Trend Micro found that the same Russian-government linked hacking group behind the infiltration of the Democratic National Committee appeared to also behind the hacking of Macron’s campaign.
Clarification: The post was published with a headline that suggested Twitter was behind the hack. It wasn’t.
This article originally appeared on Recode.net.
