UNESCO in common with other bodies that support international human rights has embraced the principle that freedom of the press is important because the media serve human rights, democracy and the public interest. Thus, press freedom requires governments not only to refrain from censorship, and harassment of the press and journalists, but also to take positive measures to support the press in its watchdog role.
As platforms transform media, and in the process undermine that function of the press, by controlling distribution, data and revenue, the question of what governments should do about platform dominance of the press is increasingly critical.
We can find the relevant principles and standards in documents such as the Windhoek Declaration which is about to celebrate its 30th birthday and the 1997 Sofia Declaration[i] and other regional declarations which had support of civil society, and documents like UN HRC GC34 as well as the standards of the Council of Europe. What these documents have in common is a set of positive obligations on states: not only to get out of the way but to positively support media. The process of updating these is under way, but in the coming period, there may be a need for new UNESCO and civil society ‘declarations’ in order to build legitimacy and a common theory of the responsibilities of the states.
The take home point in my comment is that press freedom must be asserted against control by the platforms as well as by states, and the EU is making significant progress on this.
Challenges related to disinformation
The media are changing. We don’t need to go into detail as the examples are familiar:
- We cannot simply assume that the ‘wisdom of the crowd’ will prevail in a ‘marketplace of ideas’.
- Algorithms amplify emotion not truth or the public interest.
- The business model may be the problem: large platform companies, ultimately now responsible to shareholders for value maximization are in the data business.
In Europe in the past 5 years there has been a crisis of disinformation. There are disputes about the extent of fragmentation for example, or foreign manipulation, but at the centre there is agreement that there is a problem with the ‘truth filters’ of democracies, and the extent to which they are trusted.
The Oxford Reuters study of Digital News from 2020[ii] found that globally only 38% agreed with the statement ‘I think you can trust most news most of the time’ Perhaps more strikingly, less than half agreed with the statement ‘I think you can trust most of the news I use most of the time’. Several European countries, such as France, Hungary, the UK, Greece, and Italy are among the countries where the smallest proportion (around a quarter) agree that they “can trust most news most of the time.” The trend data show a steady decline in measures of trust in the past couple of years.
Of course there are other trends, and it is not all doom and gloom. But what is clear is this is a period of fundamental shifts in political and media institutions, and this is creating problems of legitimacy for democracy. The bonds of trust in information are fraying.
This is reflected in the Democracy Action Plan of the European Union and a range of new media policy developments. Some of the ‘solutions’ can bring with them dangers for press freedom.
Self-regulatory changes by platforms and the codes of conduct
Media and policymakers have responded. The first wave was self-regulation and the development of a code of practice on disinformation coordinated by the European Commission. With an eye on freedom of expression this was very cautious and based on voluntary action by the platform companies.
The 2018 rules set out in a voluntary code a definition of harmful disinformation and a number of commitments that signatories – the platform companies – made to reduce it. This process was based on a dialogue with stakeholders aimed at demonetizing deliberate disinformation whilst using tools based on self-regulation, transparency and voluntary fact-checking.
Weakness and difficulties of self-regulation in terms of legitimacy, capture etc
Self-regulation drew some lines in the sand. Ads should be clearly labelled, and whilst anonymous posting should be possible, bots and deliberate manipulation should not.
But this model had predictable weaknesses. As a voluntary structure it depends on incentives, rather than common rules. Like all self-regulation it is subject to capture and lacks legitimacy.
The code did result in real actions – such as demonetising clear disinformation and brand safety measures. This is part of a wider process of what is called in France: ‘responsabilisation’ of platforms which includes also commitments to ‘prioritize’ trusted content, and third-party fact checking.
Such interventions may in time have a great deal of impact on press freedom because fact checking, curation and prioritization of public interest content is what the press do.
There has however been a reluctance -or inability- of the platforms to work in a timely way to remove disinformation. This is because of the nature of the task: making judgements, at scale, about subjective standards is difficult. Unlike for example intellectual property infringement, there are few economic incentives to create automated systems of effective self-regulation.
The emerging co-regulatory model.
As so often, co-regulation was required to strengthen incentives to self-regulate, but also to raise standards of due process and appeal. The Digital Services Act, a draft legislative framework for the whole EU (not the UK!) sets out clear obligations on the large ‘gatekeeper’ platforms. The legislation, published at the end of 2020, formalises some commitments made in the code of conduct and introduces new ones too. Platforms may be obliged for example to prioritize ‘authoritative’ ‘authentic’ or ‘public interest’ content, work with fact checkers and other third parties, and audit their impact on fundamental rights.
This is part of a new settlement with far reaching implications for the economic sustainability of the press and we should watch it very closely.
One of the real strengths of the new EU framework is firstly that it is driven by the value of democracy – rather than economic growth – and that it brings together competition and other aspects with content and harm.
The shift to complex co-reg comes with dangers: in terms of transparency, legitimacy and capture.
We can think of this reorganisation of complex co regulation, in which the press has to find a place, as a redistribution of power. The danger in all this is what I call ‘censorship by the blob’ by an opaque tangle of understandings between states and private companies’. If self-regulatory codes are set out by platforms under the threat of regulation, and they cover such sensitive issues as definitions of hate speech, terrorism, disinformation or misinformation, and also implemented by platforms there may be a danger that definitions of speech control could chill speech or be perceived to do so.
This is not an abstract theoretical problem but one that is being played out now around Europe, as countries have addressed a perceived “infodemic” around the Covid crisis, and around anti-vaccine messaging. Most work on disinformation in the past year has been focused on something we all have a strong incentive to agree on: a virus that kills humans which is out of control and constitutes an emergency situation. Structures and procedures for monitoring virus disinformation cannot simply be repurposed for things we absolutely don’t agree about: such as what constitutes disinformation in an electoral context. We need to develop structures and incentives to promote truth, but not a truth that serves one or other interest, or the powerful more generally. They have to be rooted in civil society or they will not be trusted.
But the code of conduct and even the DSA are really just a part of what we could call a new settlement for the press. If we want to understand the real foundations of freedom of the press – including the laws, codes and revenue flows that determine whether a news outlet is viable or not, we have also to take into account the intellectual property, competition and wider regulatory settlement.
This will take longer. It is both an opportunity and a threat for press freedom, and it does require governments and lawmakers to take a positive attitude: what regulatory settlement will support the press in this new environment, without controlling it?
In the light of all this, the key challenge is trust, transparency and legitimacy of speech governance and communications gatekeeping.
This is why procedural media governance standards are so important. All forms of speech control, particularly new forms of complex self and co-regulation should be subject to high standards such as transparency and oversight.
I should point out three developments:
1. the work of the MSI REF which is the expert group on media governance and reform of the Council of Europe. Acknowledging that we are moving to a complex, co-regulatory system we need to develop procedural standards so that the public not only trust the media, but trust the governance structures they operate within. This means working really hard to set out new standards – particularly of independence and transparency – of these new public and private governance arrangements.
2. Work on Prominence
A key part of the new regime is prominence. The platforms have developed a range of commitments to promote public interest, trusted and authentic news. But who determines who should benefit from those measures? – and under what standards of transparency. The same Council of Europe group is designing new standards for regulating prominence across broadcast, streaming and online.
3. The importance of independent regulation. This is not a new concern. It was there in the original press freedom declarations, but we need to redefine it in the new age.
Article 39.2 of the draft Digital Services Act sets out a stringent standard for new bodies with responsibility for overseeing internet regulation.
“When carrying out their tasks and exercising their powers in accordance with this Regulation, the Digital Services Coordinators shall act with complete independence. They shall remain free from any external influence, whether direct or indirect, and shall neither seek nor take instructions from any other public authority or any private party.”
This will be very hard to achieve, and it is crucial also for democracy that civil society and citizens actually have the resources and ability to participate and oversee these concerns.
One key point is that the developments in Europe are in line with and quite focused on international human rights standards. They are not in line with US first amendment standards, which underpin the operating principles of the platforms.
Perhaps, in addition to thinking about the need to convene civil society and the media around these new challenges for press freedom, it would be useful to open up more spaces for debate of these standards between EU and US, and also the UK. Democracies need to stick together on these matters.
The Windhoek declaration is 30. But it is not even the end of the beginning. Press freedom has to be prepared for change as the world changes. We are going to need some more declarations and much, much more coordination and support for civil society involvement in how the challenge of disinformation is addressed in Europe.
When UNESCO declared the first world press freedom day all those years ago it was relatively clear who the press had to be liberated from: the state. Now it is just as important to liberate them from the platforms, and from the blob of public-private censorship that is being put in place.
Dr Damian Tambini is Distinguished Policy Fellow in the Department of Media and Communication at LSE, where he also serves as Programme Director for the MSc Media and Communications (Governance).
[i] Europe Forum, 26 Years After the Sofia Declaration, organized by UNESCO with the Centre for Media Pluralism and Media Freedom of the European University Institute https://accountablejournalism.org/ethics-codes/International-Sofia
[ii] Reuters Digital News Study 2020. Main report p15