Researchers Inadvertently Discover Crypto Scam Involving 15,000 Twitter Bots

The scam is still going on.

DON'T @ ME. It's hard enough to figure out when humans are running a cryptocurrency scam. Now, it seems we have to worry about bots perpetuating them, too.

While conducting a study to figure out the best way to identify Twitter bots — accounts controlled by software, not humans — researchers from security software companyDuo Security came across a network of at least 15,000 bots working together to perpetuate a cryptocurrency scam.

The researchers plan to present their study on Wednesday at Black Hat, an information security conference in Las Vegas, NV.

BOT OR NOT. The researchers set out to figure out how to identify bots and networks of bots. They started by analyzing the activity of 88 million public Twitter accounts from May to July 2018. The researchers then used machine learning and other data science techniques to analyze more than 500 million tweets from those accounts.

This analysis led to the discovery of the crypto botnet scam. According to their study, more than 15,000 bots were working together to spread a fake crypto giveaway. The scheme requires users to pay some amount of an existing cryptocurrency and they'll get a much greater amount in return (how this would make a viable business model is certainly questionable). But instead, the scam takes their crypto but gives them nothing in return.

The researchers say they reported the scam to Twitter, but the platform has yet to shut down the thousands of accounts. A Twitter spokesperson told TechCrunch the company is currently investigating the issue.

RISE OF THE TWITTER BOTS. If you spend much time at all on Twitter, you've likely encountered at least one bot. These bots can wield a tremendous amount of power, spreading fake news and even influencing elections.

Not only is it hard to detect bots online, we're now reaching the point where detecting them over the phone is no easy task. Studies like the one the Duo Security team did can help address the issue by leading to better detection methods. But more stringent regulations, such as rules that require companies to identify online bots as bots the same way some propose forcing audio bots to identify themselves, could accomplish the same thing.

READ MORE: Duo Security Researchers’ Twitter ‘Bot or Not’ Study Unearths Crypto Botnet [TechCrunch]

More on Twitter bots: We’ve Seen What Bots Do to Democracy. Are We Adapting Fast Enough?