• Aucun résultat trouvé

16081 Twitter deletes 125,000 Isis accounts The Guardian

N/A
N/A
Protected

Academic year: 2022

Partager "16081 Twitter deletes 125,000 Isis accounts The Guardian"

Copied!
1
0
0

Texte intégral

(1)

16081 Twitter deletes 125,000 Isis accounts

The Guardian Date: 05.02.2016

Twitter has deleted more than 125,000 accounts linked to terrorists since mid-2015, the company announced, offering some of the most detailed insight yet of how Silicon Valley is collaborating with western governments in its fight against Islamic State.

The social media company is relying on a mix of human judgment and technology, developing teams of specialists in the US and Ireland that comb through thousands of suspect accounts.

Isis has also been linked to automated accounts, or “bots”, that churn out extremist rhetoric.

These automated accounts can sometimes be caught with tools normally used to fight spam.

“We condemn the use of Twitter to promote violent terrorism,” the company said in a statement on 5 February. “This type of behavior, or any violent threats, is not permitted on our service.”

Twitter and other tech companies have long policed their content for signs of Islamic extremism. Until recently they have not been willing to discuss the details of their policies or procedures, which have been viewed as controversial because of the complex and political nature of the judgments behind removing messages.

The procedures require Twitter’s specialists to make tough calls on individual tweets and accounts, including to what degree a message is euphoric, controversial or extremist.

These efforts also create risks for the company, which has frequently defined itself as a home for free speech and open debate. In 2014, then-Twitter chief executive Dick Costolo said he had received death threats over Twitter’s removal of Isis content.

But western politicians increasingly have applied pressure on US technology firms to become more of an active combatant in its fight against Islamic extremism. So far, Silicon Valley appears to be receptive.

Speaking in Davos last month at the World Economic Forum, Facebook COO Sheryl Sandberg suggested Isis could be quelled by posting anti-Isis content on Facebook. In a closed-door meeting last month in San Jose with US national security officials, she and other technology executives discussed the idea of using an algorithm to try and automatically flag Isis content.

Meanwhile, Google executives have discussed the idea of forcing Isis operatives off the public internet.

Twitter is free, international and allows anyone to sign up for an account and post what they please, which has made it an attractive recruiting tool for terrorists. In 2015, the US military killed a British national-turned-Isis-hacker Junaid Hussain partly because of how he used Twitter to plot attacks on the west.

403 words

Références

Documents relatifs

We addressed the problem of user matching across StackOverflow, Github and Twitter social networks. We proposed a methodology that combines different matching strategies and makes

Using two large hand-labelled datasets of tweets containing spam, we study the suitabil- ity of five classification algorithms and four di↵erent feature sets to the social

In this regard, I would urge the two Bloomberg Global Initiative priority countries to speed up the developments already taking place at national level to control tobacco and to

Our submission to the PAN 2019 bots and gender profiling task relies on a two-step binary classification in which we first predict whether Twitter accounts are bots or not

The model of classification of Twitter accounts in bot or human and later in male or female proposed in this work is based on a neural architecture that through convolu- tional

This is es- sentially a language modelling task and is performed using two ML models: a logistic regression on hand-crafted features (LASSO) or a pre-trained and fine-tuned

The quantitative analysis can offer an overview of the user behaviour, in terms of interaction patterns and typology of content posted and clearly illustrates the activity level of

This research is aimed at mining Arabic tweets to measure customer satisfaction toward Telecom companies in Saudi Arabia, and to predict the ratio of customer churn..