Twitter bots have become a sort of RSS feed for tweets

The new wave of witter bots is a trend I like. These basically gather tweets of different types onto one account. These are different from bots that spread misinformation and spam, which Twitter banned. Because of this, CEO Jack Dorsey mandated that legitimate creators put identifying information in their bios, such as creator handles. So, for example, you can see this screenshot of the Language Bot Twitter Bots that can be created through the official Twitter API or third-party providers.

wp 16333584063635167534062989671814
The use of specific identification methods makes its legitimacy clear.

There are Twitter bot accounts on many topics such as medical marketing and language. Cats are also a common Twitter bot focus. I have found many of them on my own and compiled some of them into a Twitter list. ( The list is not limited to Cat-themed Twitter bots, though) The confusion below is an example of the weakness of the search AI Twitter bots use. The Cat Retweeter was not accurate, as an early example shows before using a better keyword. But even the keyword like nonfeline posts, such as mention the University of Kentucky wildcats team and the singer Doja Catt I will give more examples of this later.

Imweet shws how certain features like profile pictures or even handled can trick the Twitter bot alogrithim in to picking up highly irrrelvant stuff.
the profile picture and/or handle might have caused this problem. The tweet is clearly not about cats

Tweeted initially by (@gay_cats_aaah) on March 27, 2021.

Twitter has worked to give legitimate bots a space, according to Social Media Today. Creators must register for verification of their Twitter bot accounts, like official celebrity accounts. In light of Twitter verification inaccuracies as of late, I wonder if this will be effective. In addition, Twitter still has trouble with the original impersonation bots despite its efforts to curb them. Twitter also created source labels for each of the verified bots to further show their legitimacy. For example, on Jaguars Daily, you will have jaguar-bot as the source label under the tweet.

Twitter Bot types

Twitter bots function on two algorithms. The first and seemingly more common one crawls tweets to search for specific keywords and hashtags s For example, The Language bot mentioned above looks for tweets with the#langtwt, and #100daysoflanguage hashtag. Others, such as Jaguars Daily, use a manual search algorithm in which the creators search for the desired image themselves.

Twitter Bot Algorithim errors examples

Both Twitter bots algorithm types are that they will pick up irrelevant stuff. Sometimes it is close, but others are way off, as the two examples below show. The Jaguar Daily Twitter bot creator, Luis Peres, had to tweak his algorithm a couple of times, especially after violent images \kept appearing. Since he is Brazilian, he uses the Portuguese word “onça,” an indigenous-derived word for beast. The problem is that it figures in other animals’ names. This resulted in the bot picking up other wildcats’ pictures, such as the mountain lion and the similar-looking leopard. To accommodate this, I told Luis I would not mind a name change. Still, someone else scolded me thinking I was arguing when Luis himself suggested it.

lazy kitty #JaguarJueves

An example of how similar looking subjects can end up on a Twitter Bot pages.  Comprasion image for text above
Lazy indeed but a similar-looking snow leopard. You can literally spot the difference.

Initially tweeted by Jaguars Daily 🐆 September 9, 2021.

Always be yourself. Unless you can be a jaguar, then always be a jaguar. #Caturday

Image from original Jaguars daily tweet shows that weakness of Twitter bot search algoritihims
Another example of how inexact search terms and keywords can lead to skewed results

Tweeted initially by Jaguars Daily 🐆 ) on September 11, 2021.

Still a wildcat, but the auto-generated captions look off like this above, So I wrote in response, “Why would I want to be a jaguar when I am just as powerful as a mountain lion. “

Manual search at least gives the bot creator more control over results but seems more time-consuming. Specific keywords, such as long tails m in both, will help improve accuracy. I suggest using the word in whatever language will give the most accurate results for a multilingual one. For example, Jaugaura would be more precise than Portuguese “onça.” And Bot creators should also consider a more precise search term called long tails in their search. They can also boolean terms or quotes around the search term.

How about some Fan Twitter Bots ?

Another type I most would like to see is the Backstreet Boys news bot since \Backstreet Boys lyrics Twitter bot already exist. So a possible name could be Backstreet Link

. In conclusion, I hope Twitter will continue enhancing security and identification features for them. Otherwise, it will be hard to tell the fakes from the real ones, as with fan accounts. Like lists and topics, it is an excellent way to see tweets on a specific subject from multiples sources at once

Visits: 12

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.