Automated social media accounts are being used by both sides in the Brexit debate, a new report shows, with some experts fearing that a sudden surge in activity by the "bots" could influence the referendum vote.
Researchers from Oxford University and Budapest's Corvinus University have found that bots are playing a "small but strategic role" in the social media conversation around Britain's referendum on European Union (EU) membership, the report released Monday explained.
And with Remain and Leave camps neck-and-neck in recent polls, the potential impact of bots on voting trends could prove to be significant.
Bots are automated programs that post from social media accounts of their own accord, with varying levels of human curation and, at face value, many are indiscernible from accounts managed by people.
Some bots simply deliver news or update feeds for organizations — including CNBC — while others are more complex, with an ability to interact with users and produce original content. A subset can be also be given "overtly political" or malicious tasks to "manipulate...choke off debate, and muddy political issues," particularly when public opinion is polarized, the report explained.
Bot traffic is now estimated to make up over 60 percent of all online traffic, marking a 20 percent jump from just two years ago, the report noted, citing research from data and cyber security firm Imperva.
After analyzing over 1.5 million tweets in the week from June 5 to 12 using Remain, Leave and neutral hashtags, researchers found that the most active 1 percent generate nearly a third of all content, with the top contributors to those hashtags being generated by bot accounts.
The study comes at a key moment for Britain, with the debate around Brexit reaching a fever pitch.
The most active, discernible bots in the study were relatively simple — mostly retweeting content like news stories and quotes from political leaders, rather than misinformation campaigns or attacks on opponents, Philip Howard, a professor at the University of Oxford and one of the report's authors, told CNBC.
However, Howard said there "is fear that that in last 24 to 48 hours before the vote, a big bump in bot activity could affect a percentage point or two of voters," pointing in particular to undecided voters who could theoretically help break the near-deadlock between the two campaigns.
"Public opinion research says majority of people don't make up mind on a referendum question until 48 hours before election," he highlighted.
"If someone has crafted a large bot, a bot net comprised of thousands or tens of thousands of accounts, with someone waiting to to activate them, that's a potential problem," Howard explained.
Howard said that the bots identified so far don't seem to be run by governments, but are being set up by "organizations and political leaders who are aggressively campaigning for the outcome they want." In some cases the accounts do disclose that they are bots, he explained, while others leave their natures hidden.
A spokesperson for StrongerIn, the lead remain campaign group, confirmed to CNBC on Monday that they have never sanctioned or funded the use of automated social media accounts.
Vote Leave, the official pro-Brexit campaign group, did not respond to comment requests when contacted by CNBC.
Despite the potential impact of a ramped-up social media campaign by supporters of either camp, Joseph Evans, a digital media analyst at Enders Analysis told CNBC via email that there is little appetite among regulators to place any form of electoral reporting restrictions on online and social media.
While social content has played an increasingly prominent role in electoral campaigns in recent years, with Barack Obama credited with wielding the first effective social media campaign in the 2008 election, Evans played down the ultimate influence of social media like Twitter in votes like the upcoming referendum.
"Of course social media discussion could swing undecideds, but I think the effect here is likely to be marginal, as the political speech on these platforms so often consists of trading slogans, self-congratulation and a near-total disregard for accuracy," Evans explained.
"I think the public tends to mistrust what they read on social media, although perhaps not as much as they should."