With elections looming in the UK and US, social media companies are still failing to deal with manipulation on their platforms, NATO-backed researchers have concluded.
The NATO Strategic Communication Centre of Excellence - an independent organization that advises NATO - bought social media engagement for 105 different posts across Facebook, Twitter, YouTube and Instagram over a three-month period.
In total, it spent around $333 with 11 Russian and five European companies that sell fake social media engagement, netting 3,500 comments, 25,000 likes, 20,000 views and 5,000 followers.
The team then identified the 18,739 accounts that were being used to deliver the fake engagement and reported them to the four platforms concerned - only to find that, three weeks later, 95 per cent of the accounts were still active.
"This means that malicious activity conducted by other actors using the same services and the same accounts also went unnoticed," the authors note.
Most of these accounts' activity was centered around commercial companies, although the accounts were found to engage with 721 political pages, including 52 official government profiles and the accounts of two heads of state.
Twitter, it turned out, was best at removing the fake engagement, with around half the likes and retweets eventually being removed. Facebook, by contrast, removed very little fake content, although it was the most effective of the platforms at blocking the accounts.
YouTube was the worst platform at removing fake accounts, and Instagram was the easiest and cheapest platform to manipulate, failing to suspend any of the accounts that the researchers reported.
It's presumably this failure to take action against the 'manipulation service providers' that makes them so confident: "Rather than a shadowy underworld, it is an easily accessible marketplace that most web users can reach with little effort through any search engine," the researchers write.
"In fact, manipulation service providers advertise openly on major platforms."
The authors point out that this was a pretty simple experiment to carry out, and that there's no reason the social media companies couldn't have been using similar tactics to weed out the fraudulent accounts themselves.
"Self-regulation is not working. The manipulation industry is growing year by year," they conclude. "We see no sign that it is becoming substantially more expensive or more difficult to conduct widespread social media manipulation."
Update: "While no anti-spam system will ever be perfect, our teams work very hard to manage spam views to less than one percent of all views. We have additional safeguards in place to mitigate the impact of these views on all of our systems," says a YouTube spoksperson.
"We also periodically audit and validate the views videos receive and, when appropriate, remove fraudulent views and take other actions against infringing channels."
