Triadic Closure and Instagram Bots
The topic that has fascinated me the most during these past couple of weeks was the idea of Triadic Closure. Basically, the principle states that a friend of your friend will most likely come in contact or become your friend sometime in the near future. This tactic is used by companies to suggest potential friends and followers on multiple social media sites. Because this tactic is used to suggest friends based on connection, I started wondering if I could use it to find bots based off lack of connection.
https://aimsmmarketing.com/instagram-bots-why-you-shouldn-use-them/
Here is an article on why people shouldn’t use bots and one of the first reasons was the lack of real engagement presented by bots. A need for human connection is why social media is used in the first place and why the tactic of Triadic Closure principles works so well when suggesting new friends.
https://www.avira.com/en/blog/how-to-spot-a-bot-on-social-media
This article, written by Diana Plutis, talks about how to spot bots on social media sites. Alongside profile structuring and activity frequency, Plutis discusses how networking tracing can help spot bots— a topic that was briefly covered in one of our problem sets.
- “Bots are… often part of a network,” she says and you can check friends/followers to see if they make sense within the ‘network’ they are attempted to promote. For example, if someone is big on activism, most of the people that they follow or interact with are likely to be in that same general network.
- Plutis also states that Bots befriend many people to grow their platforms and warns that “[If] you have no friends in common… it’s best to decline the friend request.” Once again, the triadic closure suggests that if two people know a person in common, those two people are very likely to have some form of connection. This can be that they are friends, they like the same content, or they have other mutual friends.
I took to Instagram and found a user that I suspected to be using bots and looked through their feed. My first indicator that this account wasn’t legitimate was the like to comment ratio. For a post with over 500 likes, there were less than 50 comments and none of the comments were interacted with, but more proof is needed. I, then went to the followings of said account and found that they only followed 37 accounts. It took a while, but I found that NONE of the 37 accounts they were following followed them back and that NONE of the 37 followed each other. I even went to see if any of the following liked or commented on their recent picture and found no connections or ties. And none of the accounts seemed to be promoting similar concepts or ideas to even be apart of a general network. This user was just an example of a node connecting to many other nodes without any of those nodes having edges to form a triangle. In the real world, not all your friends will come in contact to all your other friends, yet you can find many instances where your connection to two people creates a connection among them. In this case, I found none.