A new study of fake and human accounts on Twitter shows that the low attention span of humans distinguish them from bots.
The research was conducted by the University of Southern California who analyzed the activity patterns in two datasets of Twitter
The first was a set of ‘tweets’ correlated to the French presidential election in 2017 while the second was a collection of posts by bots and human accounts that had already been hand-labeled. The researchers took a comprehensive look at both sets of tweets and assessed the number of retweets, replies, and mentions they attracted.
Interestingly, the results showed a significant difference between humans and bots. According to the study, human interactions increased over time but they simultaneously tweeted less.
On the other hand, the bots balanced their activity and produced more content at specific times.
The observation shows that humans usually get tired after composing original content and/or get distracted by other material on the internet.
The finding was used to teach a bot-detecting algorithm called Botometer. This helped Twitter differentiate between human and fake accounts, limiting the appearance of superficial profiles on the platform.
Read next: Twitter Adds A Button To Let Users See Quoted Retweets
The research was conducted by the University of Southern California who analyzed the activity patterns in two datasets of Twitter
The first was a set of ‘tweets’ correlated to the French presidential election in 2017 while the second was a collection of posts by bots and human accounts that had already been hand-labeled. The researchers took a comprehensive look at both sets of tweets and assessed the number of retweets, replies, and mentions they attracted.
Interestingly, the results showed a significant difference between humans and bots. According to the study, human interactions increased over time but they simultaneously tweeted less.
On the other hand, the bots balanced their activity and produced more content at specific times.
The observation shows that humans usually get tired after composing original content and/or get distracted by other material on the internet.
The finding was used to teach a bot-detecting algorithm called Botometer. This helped Twitter differentiate between human and fake accounts, limiting the appearance of superficial profiles on the platform.
Read next: Twitter Adds A Button To Let Users See Quoted Retweets
