Certified company | Engineering the world

Get in touch

Little Evidence About The Influence Of Social Bots And Fake News Found

Dec 29, 2017 | Uncategorized | 0 comments

Social bots spreading fake news are far less influential than many reports suggest, according to a new analysis.As a data journalist tells, he had failed to observe the influence of smart bots in the general election.

Fake News

The election of Donald Trump in the US was also credited by many to an extremely successful campaign on the internet, supported by social bots. In Germany, the SPD candidate Martin Schulz had warned before the election that all parties should refrain from using such funds. As a new analysis is now to show, the often formulated fear of the influence of fake news, which is deliberately disseminated by automated algorithms, seems exaggerated.

As the data journalist Michael Kreil explains in his speech at the 34th Chaos Communication Congress in Leipzig, he could not have noticed any significant activity of such social bots on Twitter – they just would not exist. Influential and particularly active Twitter accounts would be operated almost exclusively by “active citizens, hashtag spammers, media and journalists” – which is joined by even bots that do not classify themselves because of the simple structure but as a sophisticated means to manipulate opinion. 

Also Read: Google’s Fight Against Fake News Begins!

Golem Kreil also finds a lot of criticism for the makers of previous studies on the subject, see the particularly active accounts and the number of tweets written as a clear feature that stands behind the account a bot. In his opinion, various studies would set arbitrarily limit values here. In his investigation of 12 accounts that were most active on the day of the US election among relevant hashtags, he had not found a single bot.

Also Read: Inventor of world wide web, Sir Tim Lee target’s fake news

Even with his extensive observation of the activity to the general election, Kreil could find no evidence of interference by social bots.Using the Twitter API and 700 access tokens provided by other users, the investigation looked at millions of accounts and tweets, but the journalist emphasizes that there are few reliable methods for identifying fake accounts. Therefore, it is now important first to review the basic assumptions of published works and to develop a new methodology. This data is provided by Kreil on Github .

Source

let’s get connected

Have a Question?

If you have any questions or need to discuss about your project
Feel free to reach out to our friendly team.