Congress is investigating how Twitter bots may have influenced the US election

It’s getting harder to tell who is real on Twitter.
It’s getting harder to tell who is real on Twitter.
Image: Reuters/Kacper Pempel
We may earn a commission from links on this page.

Twitter executives will be grilled tomorrow morning (Sept. 28) on Capitol Hill by members of the Senate Intelligence Committee, which is probing Russia’s meddling in the US presidential election.

The main focus of the questioning will be automated Twitter accounts, a.k.a. bots, and whether Twitter is doing enough to curb the ones that are spreading propaganda and misinformation, one person briefed on the committee’s preparations told Quartz.

Twitter’s bot army is so widespread that as many as 15% of its monthly active users—49 million, based on Twitter’s latest figure for its total user base—may be bots, estimated a study (pdf, p. 9) of English-speaking accounts earlier this year. Groups like Securing Democracy are attempting to track Russian propaganda on the network, by tracking “hashtags, topics and URLs promoted by Russia-linked influence networks” there.

“We’ve heard they [Twitter] don’t do a good job of removing ‘sock puppet IDs’ [Twitter users pretending to be someone else] or Russian bots,” the person briefed said. Senators are worried that bots spewing political spam or hijacking conversations result in real information getting buried, this person said. They’re also interested in what effect bots retweeting articles and information may have on Google search results. Facebook may be doing a better job of monitoring its users and customers, and removing harmful fake ones, the committee believes.

“There a universal appreciation that bots are becoming problematic on Twitter,” Robert Gorwa, a researcher with the Project on Computational Propaganda at the Oxford Internet Institute, told Quartz. “What isn’t appreciated is how much Twitter has enabled this by encouraging automation.” Twitter’s developer policies tell users how to use automation and the company provide guides on how to use bots to spread messages, he noted.

In the early days of Twitter, bots were easy to spot by their “egg” avatars, unusual ratios of following to followers, or obviously machine-written tweets. Now, as many have fleshed-out Twitter profiles or even entire online personalities (pdf, p. 16) and technology has improved, “they’ve gotten so sophisticated, we can’t recognize them,” Gorwa said.

A Twitter spokesman told Quartz the company was cooperating with the committee and confirmed it would brief it this week. “Twitter deeply respects the integrity of the election process, a cornerstone of all democracies, and will continue to strengthen our platform against bots and other forms of manipulation that violate our terms of service,” the spokesman said.