Russian efforts to influence US policies and influence in public opinion were consistent, largely in interaction with the target audience, largely, according to a report from the Oxford Computer Propaganda Project published today. Based on data provided to Congress by Facebook, Instagram, Google and Twitter, the study paints a picture of a year-long campaign that is less pleasing to businesses.
Report, which you can read hereIt was published today but was distributed to some outlets over the weekend. It summarizes the work of the Internet Research Agency, the Moscow Impact Factory on the Internet and the Dwarf Farm. The data covers different periods for different companies, but 2016 and 2017 have largely demonstrated most of the activity.
A clearer picture
If you have entered this novel from time to time in just the past two years, the Comprop report is a great way to get a comprehensive view of everything, without any "take it seriously", which makes you intersect the facts.
If you follow the story closely, the value of the report is often derived from the specifics and some new statistics from the data, presented by researchers at Oxford seven months ago for analysis. The numbers appear to have been slightly higher or more severe than those reported by the companies themselves in their voluntary reports and carefully scrutinized testimony.
Previous estimates have focused on the ambiguous measure of the "face" or "IRA" content placed on these social scales. This had a double effect to increase the affected-to- More than 100 million On Facebook alone – but "watch" can easily be underestimated; after all, how many "see" things online every day?
Oxford University researchers estimate the volume of interaction better Facebook First, with more specific and sequential numbers. For example, in 2016 and 2017, nearly 30 million people on Facebook Already shared Russian propaganda content, with a similar number of likes, millions of comments generated.
Please note that these ads are not ads paid by Russian shell companies to a specific schedule – these are pages and groups of thousands of users who have actively engaged in publishing posts, memos and misleading information on restricted news sites associated with advertising accounts.
Of course, the content itself has been carefully planned to touch on some contentious issues: immigration, arms control, interracial relations, etc. Many different groups (ie black Americans, conservatives, Muslims, and home-based communities) have been targeted. All generated a great interaction, where analysis of the statistics above shows the following:
Although the target communities were surprisingly diverse, the intent was highly focused: the lack of partisan divisions, the suppression of leftist voters and the activation of right-wing voters.
Black voters in particular were a common target across all platforms, and a great deal of content was disseminated to keep ethnic tensions high and to interfere in actual voting. Mimz has been proposed to suggest to followers blocking their voices, or with deliberately incorrect instructions on how to vote. These efforts were among the most popular and popular IRA campaigns; it is difficult to judge its effectiveness, but certainly it was accessible.
In a statement, Facebook said it was cooperating with officials and that "Congress and the intelligence community are in a better position to use the information we and others provide to determine the political motivations of representatives such as the Internet Research Agency." "I made progress in helping to prevent interference in our programs during the elections and strengthened our policies against voter repression before mid-2018 and funded independent research on the impact of social media on democracy."
Instagram in height
Based on the narrative so far, one might expect that Facebook – which has become the focus of most of it – will become the biggest platform for this propaganda, and that it will culminate in the 2016 election, when the obvious goal to help Donald Trump is made.
In fact Instagram Has received more or more content from Facebook, and has been treated on a similar scale. Previous reports revealed That about 120,000 jobs linked to the IRA on the Instagram had reached several million people in the run-up to the election. However, Oxford University researchers found that 40 accounts received a total of 185 million admissions and 4 million comments during the period covered by the data (2015-2017).
The partial interpretation of these fairly high numbers, which is contrary to the most obvious narrative, may be that the deployment of the IRA has actually increased after the elections – for all platforms, but particularly on Instagram.
Posts in the Instagram related to the IRA jumped from an average of 2,611 per month in 2016 to 5,956 in 2017; note that the figures do not match the table above exactly because the time periods vary slightly.
twitter While the number of jobs is very large, it is completely stable at less than 60,000 a month, a total of 73 million participants during the period studied. To be quite frank, this kind of teddy bear and puppet puppet activity is very common on Twitter, and the company seems to have done little to frustrate it, and hardly remember it. But there certainly were, and they often reused existing bot networks that had previously been in politics elsewhere and in other languages.
In a statement, Twitter said it had "made significant strides since 2016 to address the manipulation of our services, including the release of additional data in October relating to previously disclosed activities to enable further independent academic research and investigation."
The Google It is also difficult to find in the report, but not necessarily because it has an impact on the Russian influence on its platforms. Oxford researchers complain that Google and YouTube were not only stingy, but seem to have actively tried to thwart the analysis.
Google has chosen to provide the Senate Committee with data in a format that is not automatically readable. Evidence was provided that the Irish Republican Army purchased ads on Google as images of the ad text and in PDF format, whose pages display copies of the information already organized in the spreadsheets. This means that Google could have provided usable ad text and spreadsheets – in a standard format for automatically readable files, such as CSV or JSON, which might be useful to data scientists – but chose to convert them to images and PDFs as if the material were all printed on paper.
This forced the researchers to collect their data by citations and signals Youtube Includes. As a result, their conclusions are limited. In general, when a technology company does this, the data that can be provided will tell a story you do not want to hear.
For example, an interesting point appeared in a second report Published today, by the new knowledgeOf the 1108 video clips uploaded by IRA-linked accounts on YouTube. These videos, explained Google's statement, "were not directed at the United States or any particular segment of the population of the United States."
In fact, All But a few dozen of these videos are related to the brutality of the police and "the life of blacks", as you recall among the most popular themes on other platforms. It seems reasonable to expect this very narrow targeting to be mentioned by YouTube in some way. Unfortunately, they have been left to be discovered by a third party and give an idea of how reliable a statement from the company is. (Google did not respond immediately to the comment request).
Strongly seeking transparency
In his conclusion, researchers at the University of Oxford – Philip N. Howard, Bhart Ganesh and Dimitra Lutzio – pointed out that although Russian propaganda efforts were (and still are) effectively and well organized, the country is not alone in this.
"In 2016 and 2017, we have seen great efforts by Russia to block elections around the world, but also political parties in these countries are spreading misleading information locally," they write. "In many democracies, it is not even clear that the publication of computer propaganda is contrary to election laws."
"However, it is quite clear that the strategies and techniques used by the government eGovernment have an impact," the report continues, "and that their activities violate democratic norms of practice … Social media has become the natural infrastructure of participation and collective grievances and the coordination of civic participation , To being a computer tool for social control, manipulated by clever political advisers, available to politicians in both democracies and dictatorships.
Waiting for politicians is, as usual, a bit of a long shot. The burden is on social media providers and Internet services to create an environment in which malicious actors are less likely to thrive.
This means specifically that these companies need to incubate researchers and observers in good faith instead of freezing them in order to protect some internal processes or embarrassing problems.
"Twitter was used to provide researchers at major universities with access to many APIs, but it withdrew it and offered very little information about sampling the current APIs that researchers wonder about its utility even in basic social sciences," the researchers said. "Facebook provides a very limited API for general page analysis, but there is no API for Instagram." (We've already heard what they're thinking about Google's submissions.)
If the companies exposed to this report take these issues seriously, as they tell us time and again, perhaps they should implement some of these suggestions.