Workshop “The Europe Connection”
The European Connection
Synergies and differences amongst EU member states regarding intermediary regulation
In June 2018 the media policy lab together with the Center for Media Pluralism and Media Freedom in Florence, Italy, invited researchers and representatives of media authorities from different EU member states together with researchers to a workshop to discuss options and obstacles for the future regulation of intermediaries.
Four different working groups addressed the issues of transparency, data-access, non-discrimination and public value. The overall goal of this workshop was to identify common objectives and coordinate future steps to propose topics and possibly concrete policy steps to the European Regulators Group for Audiovisual Services (ERGA). Several EU proposals – such as the one on fairness and transparency for business users of online intermediation services – and changes in regulation, e. g. the AVMS directive, have started a debate on how to regulate and monitor intermediaries in the European Union. This workshop focused on the feasibility and practicalities of monitoring methods and necessary requirements for regulation.
The guiding questions were:
- How do we ensure transparency? What does the public need to know about Algorithmic Decision Making?
- How do we access data for monitoring intermediaries? Which auditing process is required? What information is necessary?
- How can equal access for media producers be ensured? Can the concept of non-discrimination be translated to intermediaries?
- How do we preserve and promote journalistic content of public value for the digital information sphere?
1. Working Group on Transparency
The European Audiovisual Media Service Directive is a starting point for a European debate about a common standard of information that should be required by intermediaries. To regulate intermediaries, the legislator should provide more specific definitions and answer the question from whom we need which information. The working group agreed that it is not only relevant for the public to know which data is collected by intermediaries, but also for what purpose it is obtained and used.
To achieve transparency, the group argued for a two level approach: regulators should have a broader access to data and information to perform their monitoring duties while the public should be given a generally deeper understanding of data usage, but not necessarily all technical details of it.
Information given to users should be organized with a high level of intuitive usability. Icons in the user interface could organize the access to more information so that answers to user questions can be explored when they arise and do not require extra effort or deeper exploration of the intermediaries’ self-description content.
Transparency, unquestionably important, is only the first step towards more user autonomy and safeguarding digital media pluralism - equally as crucial, is effective monitoring and safeguards regarding content to secure non-discrimination.
2. Working Group on Data Access
This working group focused on the question of how a monitoring for media pluralism and freedom from discrimination could be implemented. A core objective for designing a monitoring process for intermediaries needs to be the feasibility and practicality of it. The solution to intermediary monitoring is not a report, it is an observatory.
To secure a valid monitoring method, different sources of data are required. The following points describe possible steps to achieve such a validity of methods:
- The observatory as an API of APIs: Technically, data access could be achieved by a dynamic data gateway infrastructure, so no recurring report, but a technical interface that functions as an API (Application Programming Interface) of APIs is needed.
To ensure the validity of the data provided in the dynamic data gateway, additional external research could be a way to triangulate findings and thereby verify the data accuracy. Scraping tools like “Robin” which are used by the University of Amsterdam could provide such an additional perspective, which could point to discrepancies. One important question is how to set up such a dynamic data gateway and how to make its design feasible and sustainable for monitoring. One way to go could be a multi-stakeholder approach, by creating a board that structures and oversees the use of this gateway. This panel could be organized internationally on the EU level with members from media authorities, government, private parties and NGOs.
- In-house auditing: Another approach could be an in-house monitoring by a group of experts from regulatory and scientific institutions, led by the media authorities. The group could be provided by the intermediary with data access to conduct research over a period of time: it would thus get a deeper understanding of internal processes, but would be challenged to find ways of ensuring the accuracy of the data provided.
- Options of research for now: Even under more restricted circumstances regarding access to data, different research approaches already exist to better understand the usage of intermediaries. The most preferable one is the use of anonymized APIs (such as Netvizz) since it best meets ethical considerations and privacy standards. A more promiscuous use of APIs should be reflected in both regards and questioned regarding its necessity. Other research options are user recordings (e. g. with webrecorder.io), archiving methods, browser tools that require user consent, interface scraping (which is often blocked by services) and last but not least qualitative methods.
3. Working Group on Non-Discrimination
The issue of non-discrimination has always been a complex one for media regulators, but in the digital era it has become even more nuanced. How can personalized services be held to provide equal access? Do we need to identify new mechanisms to keep the public sphere plural and open for anyone to participate? How do we define criteria to guarantee non-discrimination?
The underlying goal of non-discrimination policies is to secure media pluralism. Therefore, the question is how to safeguard digital media pluralism by new means. The following ideas are meant to start a debate on how the goal of non-discrimination can be re-enacted in the digital realm.
- Equal treatment: Intermediaries should define and follow their own criteria for content curation in an equal way. To understand this process and to safeguard its validity, intermediaries need to create transparency regarding their own guidelines to privilege or downgrade certain types of content. Especially if political positions might be relevant for the outcome of content management - even in the most admirable ways - these decisions should be transparent and comprehensible.
- Structural equality: Another important approach to provide more diversity in content management is to look at the data sets by which the recommending algorithms of intermediaries are trained. Biases in the training data sets can be one crucial source of discrimination. Studying biases of algorithms as media authorities could also mean to look at the diversity in corporate culture and any data involved in the training of relevant algorithms.
Social media council: To support the goal of non-discrimination, intermediaries could be required to provide information about the algorithm decision-making processes that influence the content management. For the individual user, this information would be overwhelming in its complexity and even for the media authorities the data provided might be difficult to manage. Social media councils might assist in this area. These councils could be mediators between regulatory authorities and the companies in question. This approach would combine co- with self-regulation.
4. Working Group on Public Value
In the mainstream media system, public value content has a privileged position in many EU countries. There is a widespread consensus that shared high-quality media content provides an important common ground for public discourse in democracies. Now especially, public value content is crucial to counter disinformation campaigns, but only if it is found. The digital attention economy changes the rules of how media content is consumed. Therefore, the question arises as to whether there is a necessity to highlight certain content online, to re-instantiate the idea of public value in the digital sphere. The following ideas are sketching out how this goal might be achieved.
Defining public value: Firstly, to reach the goal of promoting public value content online, we need to define precisely what we mean by public value for the digital realm. Only by providing a detailed definition of public value, can regulators incentivize digital media producers to provide public value content to an online audience. An important question is who defines what public value means today (e.g. the regulator or the user?) and how a process to renegotiate its meaning could be initiated publicly.
Create incentives: To promote the production of public value content, regulators can create incentives, e. g. privileges in rankings, financial participation of license fee or other types of funding. Potential risks of direct financial support should be investigated beforehand and avoided to avert political interference.
Green light for public value: Facing an information overload, users are having difficulties in distinguishing the quality and trustworthiness of online content. It might be an option to visually assist users in differentiating between public value and non-public value content - similar to an iconic traffic light system as used by the food industry.
Gather evidence: Before regulators should take concrete steps to support public value, more information might help. Only precise empirical knowledge about the lack of information for citizens can help authorities to design steps to counter this issue.
Conclusion and Next Steps
One important conclusion drawn from this inspiring workshop was a need for more exchanges about studies and findings on media usage online to define common issues more empirically. Another was that media authorities must establish new priorities: the digital information sphere not only requires the regulation of media authorities, but it also forces them to reassess their priorities and restructure their resources to match the needs of a digital society. The media authorities must enter into the process of gaining new competences internally and focus on new analytical capacities.
Many issues regarding digital media pluralism are shared within many of the EU member states. Solutions will require steps at EU level and joint efforts to safeguard media pluralism in a shared digital world.
We would like to thank all guests again for their participation. We believe many of the ideas presented here are extremely valuable and should be discussed further in EU circles (such as the ERGA and EPRA) as well as in each EU country.