Like many, including myself, many times noted that Twitter has a problem with bot profiles.
For many years, users have been complaining about the impact of bots and fake accounts on the platform, and although fake profiles on Twitter ranged from 5% to 15% in various research reports, their presence appears to be even more significant. Researchers have repeatedly pointed out massive congestion of bot accounts that are used for malicious purposes, with the most worrisome reinforcement of political messages to drown out opposing views.
But now Twitter says that many of these statistics are incorrect – and they are not only incorrect, but also potentially dangerous for public comment.
Twitter claims that over time, it has improved its systems for detecting and deleting problematic bot profiles, and the insinuation of these reports – that Twitter allows for massive retweets on such a scale to influence public opinion – is wrong.
In a post co-written by Yoel Roth, Head of Honesty, Twitter, and Nick Pickles, Director of Global Public Policy, they describe the various limitations of third-party bot analysis on Twitter and point out their own results as a more accurate measure of performance.
According to the post:
“For several years, automated accounts have been a problem for us. We focused on this, made investments and made significant progress in addressing them on all Twitter surfaces. ”
Twitter says its Transparency Report is a good indicator of progress in this regard – in its latest update, Twitter says its total number of calls issued to suspected spam accounts – including malicious bots – was 15.4 million, spanning June 2019. .
Using this as an indicator, Twitter had 139 million mDAU at that time – therefore, based on this, as a very crude measure, the total number of possible fake profiles detected during this period could be only 11.08%. But this figure also includes all possible violations that may indicate that the actual number of the Twitter spam account, based on the actions taken by it, consists of several lower digits – and they are really and systematically detected and deleted.
Given that this number also improves over time (as shown in the graph), Twitter says it basically gets rid of deceiving bots. Which leads to another key difference Twitter wants to make.
“It is important to note that not all forms of automation are necessarily a violation of the Twitter Rules. We saw innovative and creative ways to automate the enrichment of Twitter — for example, accounts like @pentametron and @tinycarebot. ”
Twitter also notes that business profiles can also use bots and automation to improve customer interactions, so you cannot mark all bots as bad, which further obscures the numbers.
So, which bots, according to Twitter, violate his rules?
Twitter lays out various types of misuse, which will lead to the prohibition of the profile of the bot.
Malicious use of automation to undermine and disrupt a public conversation, for example, an attempt to force something to change
Artificially reinforcing conversations on Twitter, including by creating multiple or overlapping accounts
Creation, extortion or purchase of fake obligations
Engaging in massive or aggressive tweets, engaging or tracking
Using hashtags with spam, including using unrelated hashtags in a tweet
Thus, the main focus of bot criticism – manipulating conversations – is really a violation, and Twitter will remove bots that violate this rule. While he discovers them, what he says, for the most part, he does.
“Our technological capabilities for proactively identifying and eliminating these violations in our service are more complex than ever. We constantly block millions of accounts every month that are automatic or spam, and we do this before they ever reach the eyeball in the timeline or on Twitter. “.
So Twitter removes the bots, and its advanced detection processes suggest that the problem is not as obvious as some reports suggested.
Whether it sounds believable with you or not depends a lot on the opinion. Or data presented in academic research on abusive tweets.
Buy Now: buy 2000 Instagram followers
This is where it’s becoming increasingly difficult to accept the Twitter counterargument – the studies they link to actually contain tweets as data, they show that tweets are often repeated word for word, and republished in quick succession during periods of popular discussion. , We know this is happening, but according to Twitter, either real people tweet it or it’s a misinterpretation of a broader trend.
Twitter emphasizes that tools like and Bot Sentinel, which are aimed at determining bot activity based on a number of parameters, are usually not accurate.
“These tools start with the person looking at the account – often using the same publicly available account information that you can see on Twitter – and determining the characteristics that make it a bot. Therefore, in essence, the name of the account, the level of twittering, the location in the bio, the hashtags used, etc. This is an extremely limited approach. “
Again, Twitter says that its own detection systems are now much more developed than these tools take into account, and therefore the numbers indicated by such tools are not only incorrect, but can also damage the discussion.
Thus, Twitter says that reports that indicate widespread manipulation of bots, including those referenced here, here, here and here, are likely to be misleading regarding the actual impact of such actions. This does not mean that Twitter has all aspects, and that it is completely free from manipulation with bots. But when people launch a profile through a bot verification tool, this is probably not enough to point a finger and blame any user for coordinated manipulations.
Based on the available data, this may be true, but it is very difficult to say – because no one knows for sure how many Twitter profiles really have bots. According to Twitter, many of the alleged actions of bots may well be real people who simply share similar messages or are actively involved in certain topics. Some alleged bot campaigns may also be flagged as such by people who simply do not want to accept actual results or trends, so they reject them as artificial.
Twitter, of course, understands this best of all, having access to raw data, and, according to Twitter, most of the bot’s activity is being addressed. It is also worth noting that Twitter is considering new tags for bot accounts, such as proven ticks for bots, which can help better understand the origin of certain trends.
Again, the way you take it will come down to your personal feelings, but the numbers show that Twitter is working on a solution to its problem with bots – and that the problem itself may not be as obvious as it might seem.
Buy Now : buy 1000 Instagram likes