The head of Twitter Ireland has said there is no evidence on the platform of concerted spam activity aimed at the Eighth Amendment referendum.
“We are vigilant when there is any major event globally, whether it is the World Cup, the Oscars, or a referendum,” Twitter Ireland managing director Sinead McSweeney said. “There are always measures in place to monitor Twitter traffic but also follower patterns around any event. We have looked at and reviewed the activity here in Ireland on numerous occasions in the last few weeks,and we have not seen any evidence of concerted spam activity directed towards the referendum.”
Concerns have been raised that social media and online advertising could be used to sway the results of the upcoming referendum on May 25th. Twitter Ireland has not accepted any advertising relating to either side of the campaign because it does not accept advertisements on abortion advocacy or services.
That has left the platform outside the recent controversy on rising online ad spend that led to Facebook and Google both tightening up restrictions on referendum ads. Facebook said on Tuesday it would reject ads on the topic that were paid for by foreign advertisers, while Google on Wednesday announced it would no longer accept any advertising around the Eighth Amendment referendum.
Twitter does not have a dedicated reporting system for flagging bot accounts, but such accounts would sometimes fall under its spam category.
“Spam has always been a particular challenge for Twitter,” Ms McSweeney said. “We have dedicated teams who work on this 24/7, but...they need to constantly change their tactics because the tactics of those who seek to introduce spam to the platform in whatever form will change.”
She said Twitter’s system captured 3.2 million suspicious accounts each week globally.
“That’s an ongoing piece of work. That is supplemented by dedicated attention and review around high profile events on the platform whether political, entertainment or sporting,” she said.
She noted that Twitter Ireland had observed an issue in recent weeks with the sign up process, where users were rushing through it and following suggested accounts but skipping steps such as adding a profile picture.
“So for a period of time they look really spammy,” she said. “We have reviewed all of those accounts and we see nothing to suggest they are anything other than real people.”
She rejected the idea that the company was complacent. “We don’t rest on our laurels there because we know that malicious automation or efforts to impact whether it’s trending topics or the nature and quality of conversation on Twitter are all things that are against our rules but which people try. So we have to keep upskilling and fine tuning our tools.”