Twitter has provided updated details on its investigation into Russian election interference on its platform in 2016. Its identification of more than 13,000 more Russian-linked bots that made election-related tweets puts the total over 50,000. In addition, about 3,800 (up 1,000 from Twitter’s data in the fall) were associated with the now-notorious Internet Research Agency.
Still, Twitter denied that these accounts were a significant problem:
The results of this supplemental analysis are consistent with the results of our previous work: automated election-related content associated with Russian signals represented a very small fraction of the overall activity on Twitter in the ten-week period preceding the 2016 election.
As if to demonstrate the different scales at work here, the Twitter blog post then changed topics to its efforts to block bots and suspicious activity platform-wide.
For reference, those 3,800 IRA bots tweeted about 176,000 times during that 10-week period, of which less than 15,000 were election-related. And 677,775 people saw, followed or retweeted one of these accounts during that same period, and are being notified.
But in a way that’s just a drop in the bucket.
“In December 2017, our systems identified and challenged more than 6.4 million suspicious accounts globally per week,” the company wrote. “Since June 2017, we’ve removed more than 220,000 applications in violation of our rules, collectively responsible for more than 2.2 billion low-quality Tweets.”
It’s not exactly apples to apples, but it is a good reminder that this was more of an experiment in influence, not a full-scale push. If simple spammers can create and promote bots by the tens of thousands, Russian intelligence could easily have brought more to bear here.
Of course, the numbers are much higher on Facebook — some 150 million people are estimated to have been reached by troll accounts there.
Lastly, Twitter explains some concrete steps it’s taking to make the 2018 elections a bit less susceptible to this type of interference, which, while not actually too grand in scale, was certainly more widespread than expected.
Specifically, the company is working on verifying all candidates, escalating issues of impersonation or hijacking, and monitoring election-related conversations closely for evidence of manipulation or bot participation.