A Data analyst firm Roman origin named, Ghost Data reported their latest study suggests that around 95 million of the one billion users of Instagram is a bot or a computer program. This research is not the first time. In 2015, Ghost Data also used the same methodology.
As a result, 7.9 percent of Instagram users who were then broke through 300 million users, were bots . The findings came after researchers bought 200,000 bot accounts to analyze their behavior. Then they match the bot behavior they bought with a million other accounts.
In conclusion they can, 200,000 bots were very active than followers or other followers on the popular Instagram account. Ghost Data finds millions of accounts that have similar behavior in many popular Instagram accounts, such as accounts of programmers or famous personalities.
The bot accounts also send postings like “normal” accounts, but they usually contain photos of models commonly found on other sites. According to one of the founders of Ghost Data, Andrea Stroppa says if the current bot maker is getting smarter, even when Instagram tries wipe them out.
When the first one bot to follow about 40 accounts, each account that follows (following) bot account. Now they have reduced the ratio by about five to one. “For a long time, social networks have not been able to find bots, I think they (social networks) are better now, but they can do better,” Stroppa explained.
The Information, as quoted from Zero Hedge, Friday (20/07/2018), also found data similar to the invention of Ghost Data. The data comes from Dovetail, a startup that analyzes bots for marketing. Dovetail co-founder Mike Schimdt said that Dovetail’s methodology uses machine learning (ML).
ML’s earnings accounts do not label these accounts as bots , but behave the same. But Schmidt says if the percentage of bot accounts in Instagram is more or less the same as the estimates found by Ghost Data. Researchers who study bot activity say if false and propaganda stories spread during the 2016 US presidential election, mostly news articles, pictures and videos.
The content plays a large role, because the messages in the image as in the meme, are difficult to trace and identify. ” Bots are also great for spreading,” said Giovanni Ciampaglia, a researcher at the Network Science Institute at Indiana University.
He says if false news and political propaganda are often created by real humans, not bots , to control social media accounts. Then, the bot account plays a role to spread the false news and propaganda content.
Millions of bots like those found by Ghost Data may be held by one person and request “retweet” or “like”, as many lie or propaganda stories do. What about Facebook? Meanwhile, the problem of bot on the parent platform Instagram, Facebook is quite difficult to identify.
Researchers say tracking bots on Facebook is more difficult due to a privacy protection policy that prevents outsiders like researchers to analyze accounts. While Instagram does not have the same rules. However, unidentified (misclassified) or undesirable accounts on Facebook increased in the third quarter of 2017 to 3 percent.
Facebook that overshadowed Instagram claims to have hired 10,000 people to scan bots and malicious behavior on its social network. They plan to double the number of workers next year plus matching-learning techniques to identify fake accounts.