Study findings

• Author: Jordan Hill - Malin Carlberg - Richard Procee - Iva Plasilova - Marion Goubet
• Pages: 32-74

Assessment of th e im plem en tat ion o f t he Code of Pract ice on Disinform at ion

32

Dir ect or at e-Gen eral for Comm uni catio ns Net wor ks, Co nten t an d Tech nolog y

CHAPTER 5: STUDY FINDINGS

This section presents the findings based on the analysis carried out by the research team

following the methodology established in the evaluation plan annexed to this report.

Firstly, some overall study findings are presented, followed by an assessment of the

effectiveness of each pillar and lastly, the findings for the other evaluation criteria.

This section provides a general overview of stakeholders’ opinions regarding the Code

of Practice on disinformation and its implementation. With the exception of the monthly

and annual reporting published by the Signatories, there are very few written sources

assessing the implementation of the Code. In line with the evaluation plan, the

information presented below is based on semi-structured interviews with stakeholders,

on the two online surveys deployed for this study, and on evidence and material gathered

for the four case studies.

Overall findings

The fact that the Code exists (as a first initiative to combat disinformation) was

identified as an achievement in its own right by all stakeholders, except for two outliers

– one civil society and one business association. However, the views on the extent of

implementation of the Code and its effectiveness are diverse.

Some of the Signatories and non-Signatory stakeholders credit the Code with pushing

them to be more proactive in the area of disinformation as well as in monitoring and

reporting on this problem more closely. From the Signatory associations perspective, the

Code has contributed to a greater awareness of the issue among their members and

more initiatives to combat the phenomenon.

Nonetheless, at times it was pointed out that while the behaviour of platforms is slowly

changing, this change is not necessarily the sole result of the implementation of the Code

but rather the result of a number of initiatives, including the Code but also including

the increased dialogue and communication on disinformation in general on the national,

European and global levels. While the Code might not have been the initiator of these

debates, it undeniably contributed and continues to contribute to them.

Several respondents from across the various stakeholder groups mentioned that there

is a need for more/better definitions in the area of disinformation. Guiding principles,

for example, on what should constitute an inauthentic behaviour or a trusted information

source, should be developed. Here, the stakeholders, both from Signatory and non-

Signatory parties, point out that the European Commission should take the leading role

and facilitate the discussion and coordination among the wide group of stakeholders on

this topic.

Similarly, the Signatories as well stakeholders in general indicated that it would be useful

if a consensus across the Member States on what disinformation is could be

reached. Currently there are several varying definitions/descriptions that need to be

followed and that may be contradictory to each other. This hampers a unified approach

Assessment of th e im plem en tat ion o f t he Code of Pract ice on Disinform at ion

33

Dir ect or at e-Gen eral for Comm uni catio ns Net wor ks, Co nten t an d Tech nolog y

to the issue across the EU. Such initiative should, again, be led by the Commission by

providing a common discussion forum for the MS.

When evaluating the Code, a first complicating factor when carrying out a detailed

assessment is that the platforms’ self-assessment reports are not harmonised and

further perceived by non-Signatory stakeholders as not user-friendly. This makes it

difficult to compare the pre-Code and post-Code environments. However, going back to

the fact that coordinated disinformation campaigns are still largely prevalent, it is not

possible to say if consumers are more or less exposed to disinformation now than before

the Code was established.65

Regarding the reporting commitment, it should be highlighted that the Signatories,

particularly the platforms, went above and beyond of what was outlined in the Code. To

complement the roadmaps of the Signatories on how they intend to implement the Code,

a baseline report was produced by them in January 2019 to allow for comparison of the

pre-Code and post-Code results and observe the development of effectiveness of the

Signatories’ policies. On top of these, and in light of the then upcoming elections to the

European Parliament, the European Commission asked the Signatory platforms to

provide monthly reports on their effort to combat disinformation across the EU in the

period leading up to the elections (i.e. January to May 2019).

Overall, these reporting requirements were met by the Signatories in terms of submission

of the reports. While many highlighted the additional burden created by all this reporting

and the monitoring it required, a few also mentioned that it helped them with taking stock

of their policies and to see where further improvements are needed.

From the non-Signatory stakeholders’ point of view, the annual reports from the

platforms are regarded as lacking enough factual details to be used to their full

potential. They do not, for example, provide any indication of the quality of initiatives

undertaken since signing the Code, or provide enough indicators regarding their

effectiveness, or the platforms’ motivations for any policy changes. Besides having a

common reporting template and an agreed set of indicators to be provided, no

additional suggestions on how to improve these reports and their readability were

made.

To this end, the already ongoing dialogue between the Signatories, Commission, and

Sounding Board should continue. As mentioned by Signatory trade associations and

Sounding Board members, the dialogue should be opened up to even more stakeholders

from different areas, A careful consideration should be given to the way this dialogue

would be structured to ensure a constructive cooperation.

65 For instance: In the run-up to the 2019 European elections (March-May), Avaaz reported a total of almost 700 suspect

pages and groups to Facebook, which were followed by over 35 million people and generated over 76 million

"interactions" (comments, likes, shares). Facebook has taken down 132 of the pages and groups reported, accounting

for almost 30% of all interactions across the reported networks. Together the pages taken down reached 762 million

estimated views over 3 months. Source:

https://avaazimages.avaaz.org/Avaaz%20Report%20Network%20Deception%2020190522.pdf?slideshow.

Assessment of th e im plem en tat ion o f t he Code of Pract ice on Disinform at ion

34

Dir ect or at e-Gen eral for Comm uni catio ns Net wor ks, Co nten t an d Tech nolog y

With regards to areas for improvement, the following suggestions were raised:

• The Code should improve its common approach to terminology, interpretation of

commitments, monitoring and reporting of the implementation; have clearer

commitments with few of them being more targeted towards individual

Signatories and; develop measurable KPIs as the current shortcomings make it

difficult to monitor and evaluate the process.

• There is also no mechanism for action in case of non-compliance, besides the

annual review by the Signatories and the independent assessment by a third

party. Therefore, a mechanism for action in case of non-compliance of the Code’s

Pillars could be considered.

• The Code is an approach “a la carte”, as the Signatories can choose to sign up

for some commitments and not for the others. Because of this, the number of

commitments could be expanded to include more areas and to provide the

possibility of more targeted commitments or innovative approaches. These

should also be discussed with other platforms and associations that have not

signed the Code yet to provide them with more incentives to join the Code as

they could suggest commitments aligned with their business models.

• A regular communication (such as a newsletter) could be set up, sending short

updates to all Signatories on the progress of platforms to implement

commitments of the Code or on the work done around the Code as well as

distributed to a larger audience connected to the issue to increase the

transparency and collaboration within the Code.

• The Commission should evaluate what more could be done to attract further

players among the Code’s target groups. For example, it should be considered

whether big advertisers should be invited to sign the Code to strengthen its

legitimacy, even if they are more likely to already have internal guidelines to avoid

disinformation.

The main draw for consumers of social media to a platform is the range of voices and

the authenticity that comes with a platform designed to encourage free expression,

however this leads to a dilemma. Namely, any attempt to clamp down on disinformation

on social media is likely to restrict the so-much appreciated breadth and vibrancy of

debate potentially leading to conflicts with freedom of expression. It should be noted that

the Code of Practice already acknowledges this: ‘The aim of the Code of Practice is for

the Signatories to actively contribute to tackling the spread of disinformation while at the

same time maintaining the right to freedom of expression and open Internet (2.4 Aim and

Scope of the Code)’ and that freedom of expression is not an absolute right. However,

the study team notes that it is still crucial for social platforms to find the correct balance

between freedom of expression on the one hand and fighting disinformation on the other,

if they want to maintain a sustainable business model for the future.

In order to fully comprehend the information supplied in the two surveys, (National

Regulatory Authorities and one with the expert public), it would be useful to have a view

of what the general public, in their role as consumers, thinks of the implementation of the