UNESCO Guidance for regulating digital platforms: a multistakeholder approach 

Regulations and Policies

Draft 1.1  

December 2022

Source: UNESCO

Disclaimer: This document presents the overall UNESCO initiative and an initial draft proposal  of the guidance document for regulating digital platforms. A further developed draft of this  document will be circulated by the end of January 2023 ahead of the Global Conference  Internet for Trust, which will provide a space for debate about the broader issues behind the  paper, the proposals themselves, and future actions. A glossary of key concepts, based on  international human rights law, UN system resolutions and declarations, other soft law  documents produced by the Universal System of Human Rights, is being developed and will be  added in the next version of this draft. 

All comments should be sent to the e-mail: internetconference@unesco.org, mentioning the  specific paragraph number the comment refers to.  

Introduction  

1. UNESCO is developing, through multistakeholder consultations, a guidance document for actors seeking to regulate, co-regulate and self-regulate digital platforms, with the aim of supporting freedom of expression and the availability of accurate and reliable information in the public sphere, while dealing with content that potentially damages human rights and democracy. The scope of this guidance covers digital platform services that can disseminate users’ content to the wider public, including social media networks, search engines and content sharing platforms. While this guidance is developed for those platforms whose services have the largest size and reach, minimum safety requirements should be applied to all platform service companies regardless of size. 

2. This document aims to provide high-level guidance forthose Member States and other  relevant stakeholders that are considering how to regulate online content. It sets standards to help them in the development of legislation and policies that are consistent with international human rights standards, and which enhance the availability of accurate and reliable information in the public sphere. It also intends to serve as a guidance for co-regulatory and self-regulatory processes, as well as a concrete tool for a process of checks-and-balances, through which companies, civil society organizations, academics, the technical community, and journalists can hold accountable the players in charge of regulating, co-regulating and self-regulating this space. 

Why UNESCO? 

3. UNESCO has a global mandate to promote the free flow of ideas by word and image. As part of the Organization’s Medium-Term Strategy for 2022-2029 (41 C/4), Strategic Objective 3 is to build inclusive, just and peaceful societies by promoting freedom of expression, cultural diversity, education for global citizenship and protecting heritage. Strategic Objective 4 is to foster a technological environment in the service of humankind through the development and dissemination of knowledge and skills and the development of ethical standards. 

4. The development of guidance for Members States (including a diverse range of public entities, among which may be different types of independent regulators) and digital platform services themselves, to secure information as a public good, contributes to all five of UNESCO’s functions as a laboratory of ideas, a clearing house, a standard setter, a catalyst and motor for international cooperation, and a capacity-builder. 

5. More specifically, the development of guidance for the regulation of digital platform services builds on the Organization’s work in the domain of broadcast regulation developed over several decades. 

6. This guidance for regulation of digital platform services focuses on the structures and processes to help users have a safer, critical, and self-determined interaction with online content, dealing with content that is potentially damaging democracy and human rights, while supporting freedom of expression and the availability of accurate and reliable information in the public sphere. 

7. This guidance will: 

7.1 Take forward the Windhoek+30 Declaration on Information as a Public Good, as it calls on all parties to mainstream media and information literacy, as well as promoting increased transparency of relevant technology companies and media viability, principles unanimously endorsed by UNESCO’s Members States during its 41st session of its General Conference. 

7.2 Create a multistakeholder global shared space for the debates on regulation, co-regulation and self-regulation of digital platform services, through an inclusive consultative process and research ahead of the conference. 

7.3 Enable a network of regulators and regulatory systems to draw upon this guidance, and facilitate the creation of an international community of practice, capable of exchanging good practices on how to approach regulation  of digital platform services to secure information as a public good while  protecting freedom of expression and other human rights.  

7.4 Serve as an advocacy and accountability tool for all the relevant  stakeholders, who will be able to advocate for smart regulation aligned with  human rights, where it is missing, and to hold relevant players (parliaments,  regulators, companies) accountable, guaranteeing that any regulatory, coregulatory and self-regulatory measures discussed and implemented are in  line with international human rights standards.  

7.5 Offer inputs to “Our Common Agenda”, including the Global Digital  Compact and the UN Summit of the Future to be held in September 2024.  •

7.6 Feed into discussions about the upcoming 20-year review in 2025 of the World Summit on the Information Society (WSIS) and the review of the  Internet Governance Forum (IGF). 

7.7 Build on and gain insights from the work linked to the development and  implementation of the UNESCO Recommendation on the Ethics of Artificial  Intelligence, adopted by the UNESCO General Conference in November 2021,  and particularly regarding its guidance on digital platforms and online media  under the Policy Area for Communication and Information. 

Independent Regulation  

8. Online content represents a new regulatory challenge that many actors, including  states, are struggling to deal with. Existing regulatory systems vary from country to  country. In some jurisdictions, there may be an existing broadcast regulator which is  being granted new powers over digital platform regulation. In other states a new  regulator may be established to regulate online content. There are other cases in  which more than one regulatory body or institution oversees these issues. taking into  account the wider implications of digital content for our societies. For instance, there  are contexts where we have audio-visual, electoral, telecom, data protection  regulatorsthat deal with different aspects of the digital platform’s services. This is why  this text is using the concept of a regulatory system. Whichever is the case, this  guidance outlines the importance of establishing the independence of the regulatory  system, however constituted, as well as ensuring that regulators have the necessary  skills.  

9. In 2006, the World Bank published its Handbook for Evaluating Infrastructure Regulatory System 1World Bank Handbook for Evaluating Infrastructure Regulatory Systems, p.50  http://elibrary.worldbank.org/doi/book/10.1596/978-0-8213-6579-3 in which it says the following about independent regulation: 

9.1 “The key characteristic of the independent regulator model is decision-making  independence. This means that the regulator’s decisions are made without the  prior approval of any other government entity, and no entity other than a court or  a pre-established appellate panel can overrule the regulator’s decisions. The  institutional building blocks for decision-making independence are: organizational  independence (organizationally separate from existing ministries and  departments), financial in-dependence (an earmarked, secure, and adequate  source of funding), and management independence (autonomy over internal administration and protection from dismissal without due cause).” 

10. In a guiding document commissioned by UNESCO (2016), the expert on broadcasting  independent regulatory systems, Eve Salomon, highlighted: 

10.1 “An independent authority (that is, one which has its powers and  responsibilities set out in an instrument of public law and is empowered to  manage its own resources, and whose members are appointed in an independent  manner and protected by law against unwarranted dismissal) is better placed to  act impartially in the public interest and to avoid undue influence from political  or industry interests. This ability to operate impartially is vital to protect freedom  of expression, which is necessary in a functioning democracy. Independence is  also required for the proper operation of all of the major functions of  broadcasting regulation, including licensing, applying content standards and  positive content obligations, and ownership and competition regulation.”2https://unesdoc.unesco.org/ark:/48223/pf0000246055 see also See also Principle 17 of Declaration of  Principles on Freedom of Expression and Access to Information in Africa  

11. This guidance is principle-based – with the regulator (or the regulatory system when there is more than one regulatory entity or body), setting the overall goal for  regulation which the digital platform services must fulfil.  

12. Importantly, this guidance recommends that any regulatory system focuses on the  structures and processes that services use to make content available, rather than  seeking to intervene in actual content decisions. 

13. In addition to setting out the primary regulatory goal which is to support freedom of  expression and the availability of accurate and reliable information in the public  sphere while dealing with content that damages human rights and democracy, the  guidance suggests that the regulator or the regulatory system could specify a number  of issues that the digital platform services should address when reporting. These are currently set out in several separate sub-items –.310 in total, although this may change as the result of further consultations   The guidance goes on to set out the constitution, powers, and scope of the regulatory system.   

The benefits of this guidance  

14. The guidance should be helpful to a range of stakeholders: for policymakers in identifying some of the stakeholders, objectives, principles, processes and procedures that might be  considered in legislation; for regulators in implementation; for companies in their policies  and practices; and for other stakeholders in their advocacy and accountability efforts.  

15. Specifically, this guidance will: 

15.1 provide guidance in developing regulation that can help Member States  address moderation and curation processes of content that potentially damages  democratic discourses and structures and human rights, while protecting freedom of  expression and other human rights; 

15.2 provide guidance on the strengthening and constitution of an independent  regulator(s) or regulatory system; 

15.3 facilitate the ability of digital platforms to align to a common framework and  develop coherent systems across regions to minimise internet fragmentation to protect freedom of expression and enhance the availability of accurate and reliable  information in the public sphere  

15.4 help develop regulatory stability and a more coherent global governance  (which would benefit the development of a more diverse range of companies and local  economies) and help Member States be “future ready” anticipating new challenges;  

15.5 support platforms by providing practical and implementable guidance with a  view to realizing the regulatory goal.  

The proposed Guidance for regulating digital platforms: a  multistakeholder approach  

Section One – The goal of regulation  

16. The objective of this guidance is to protect freedom of expression and enhance the  availability of accurate and reliable information in the public sphere, while dealing with  content that potentially damages human rights and democracy.  

17. The guidance sets out how the regulatory system can oversee the conduct of digital  platform services in respect to content issues.  

18. The guidance also outlines government responsibilities to be transparent and  accountable about the requirements they place upon digital platform services,  particularly regarding the alignment with international human rights standards. For  example, governments should be open, clear, and specific about the type and volume of  requests they make to companies to remove and block content. In the case of sensitivities  about publicising these requests – for instance, content relevant to national security or  the prevention of serious crime – then the regulator or regulatory system should be able  to examine, based on human rights standards, the validity of such requests and be able  to report publicly on their findings and actions. The regulator should also be able to  scrutinize the scope of requests to ensure adequate balance between illegality and  freedom of expression. 

19. This will also require finding a means to deal with the potentially harmful content that may damage democracy and human rights – current examples include hatred of defined  groups; incitement to violence; harassment; mis- and disinformation; and hostility  directed at women, racial and minorities, human rights defenders or vulnerable groups –  while protecting international standards of freedom of expression. But we should  recognise that new dangers may arise that are not foreseen now, and that any regulation  to protect human rights must be flexible enough to adapt to new or changing  circumstances.  

20. Finally, this guidance shows that this goal can be achieved only if there is a cooperation  among the companies providing services and the regulatory systems, while being effective and implementable and providing real accountability.  

21. This guidance for regulation will be based on five key principles: platform, policies and  operations need to be (1) human rights-based, (2) transparent, (3) empowering,  (4) accountable and (5) verifiable, to help ensure:  

21.1 Platforms have content governance policies and practices consistent with  human rights standards, implemented algorithmically or through human means(with  adequate protection for the well-being of human moderators);  

21.2 Platforms are transparent, being open about how they operate (taking into  account commercial confidentiality) with policies being explainable;  

21.3 Platforms empower users to use digital services in a self-determined and  empowering manner, including being able to assess the quality of information  received; 

21.4 Platforms are accountable to users, the public, and regulators in  implementing terms of service and content policies, including giving users rights of  redress against content-related decisions; 

21.5 There is independent oversight and assessment of the impact that regulation  has on companies’ rules and practices, with a view to adjusting regulation to more  effectively protect information as a public good. 

Section Two – Fulfilling the goal 

22. Before setting out the responsibilities of digital platform services in respect to the  regulator, it is helpful to set out the responsibilities of governments that are considering  legislation to regulate processes impacting content moderation and/or curation, so that  such legislation fulfils the regulatory goals of providing and ensuring information as a  public good, while protecting freedom of expression. 

23. Governments should: 

23.1 Protect and respect users’ rights to freedom of expression, the right to  information, equality and non-discrimination; 

23.2 Respect the requirements of Article 19(3) of the International Covenant on  Civil and Political Rights (ICCPR), in that any restrictions applied to content should have  a basis in law, have a legitimate aim, and be necessary and proportional to the harm  that is being restricted; 

23.3 Ensure that any restrictions are also consistent with Article 20 of the ICCPR;41. Any propaganda for war shall be prohibited by law. 2. Any advocacy of national, racial or religious hatred that constitutes incitement to discrimination, hostility or violence shall be prohibited by law.”  

23.4 Be transparent about the requests they make to companies to remove or  restrict content, and be able to demonstrate how this is consistent with Article 19 of  the ICCPR; 

23.5 Guarantee that any content removals are subjected to the adequate due  process of law, including independent judicial review; 

23.6 Not impose indirect restrictions to companies (for example, internet  shutdowns) for alleged or potential breaches of regulations; 

23.7 Not subject staff of companies to criminal penalties for an alleged or potential  breach of regulations, as this will have a chilling effect on freedom of expression;  • 23.8 Regulators with responsibilities in this area should be structured as  independent regulators, with the proper accountability systems in place. 

24. In turn, this guidance recommends that the regulatory system expect digital platform services to have in place structures and processes and to report to them on the  following issues: 

25. Transparency of process.

How digital platform services fulfil the principles of  transparency, explicability, and reporting against what they say they do in their terms  and conditions (T&Cs) and community standards. This should include:  

25.1 Information about the reasons behind any restrictions imposed in  relation to the use of their service being publicly available in an easily  accessible format in their terms and conditions;  

25.2 How content is managed, including through algorithmic decision making and human review, as well as content that is being removed or blocked  under either T&Cs or pursuant to government demands/requests, and information relevant to complaints about the removal, blocking, or refusal to  block content;  

25.3 Any information about processes used by the platform to enforce their  T&Cs and sanction users, as well as government demands/requests for content  removal, restriction, or promotion; 

25.4 Any safeguards applied in relation to any content moderation that are  put in place to safeguard freedom of expression and the right to information,  including in response to government demands/requests, particularly in  relation to matters of public interest, so as to ensure a plurality of views and  opinions; 

25.5 How users can access the complaints process; 

25.6 Any use made of automated means for the purpose of content  moderation, including a specification of the role of the automated means in  the review process and any indicators of the benefits and limitations of the  automated means in fulfilling those purposes. 

26. Content management policies.  

The content management policies of digital platform services should be consistent  with the obligations of corporations under the UN Guiding Principles for Business and  Human Rights, the International Covenant for Civil and Political Rights and relevant  regional treaties. They should also follow best practices as expressed, for example, in  the Santa Clara Principles.5Any restrictions upon content being posted, or content removed, should be defined in law,  have a legitimate purpose and be necessary in a democratic society and be applied  proportionally.  

27. Any restriction upon content posted should be clearly set out in the platform rules,  which should be implemented consistently, without arbitrary distinctions made between types of content or between users. 

27.1 Platforms should, in policy and practice, through adequately trained  and staffed personnel, ensure that, at a minimum, there is quick and decisive  action against child sexual abuse materials, promotion of terrorism, promotion  of genocide, clear threats of violence, gender-based violence and incitement  to hatred based on protected characteristics.  

27.2 There is often a tension between national laws and international  human rights standards, which poses a challenge for any attempt to define   

global guidance on regulation. Should illegal content be defined in a  jurisdiction that may violate international human rights law, the platform will  be expected to report on how it responds to such requests.  

28. In addition, platforms should report on systems they have in place that would help  enable them to identify the following, while protecting the right to privacy and  anonymity:  

28.1 multiple accounts created by the same source;  

28.2 [false or inauthentic behaviours that promote mis / disinformation or  other damaging content];6This item was highlighted by different stakeholders consulted as requiring extra safeguards for the balancing of rights, therefore we would appreciate detailed comments on how to address this point.     

28.3 [synthetic content designed to mislead or create a false impression  (unless clearly identified as such for artistic or creative purposes)];7This item was highlighted by different stakeholders consulted as requiring extra safeguards for the balancing of rights, therefore we would appreciate detailed comments on how to address this point.     

28.4 the use of automated programmes designed to mimic users (bots); •

28.5 content created by accounts registered by state actors or otherwise  credibly determined to be state-affiliated. 

29. Platforms should then have explicit processesto deal with these phenomena, whether  it is to label and identify such content or accounts, while protecting the right to privacy  and anonymity, to restrict the virality of content arising from such accounts, or to flag  with a warning that the nature of this content could be misleading or otherwise  problematic (similar to splash page warnings now provided by banks before allowing  transactions). The purpose for these provisions is to allow users to understand the  nature and origin of questionable content and accounts and allow them to make their  own judgement as to their provenance.  

30. Finally, platforms should notify users when their content is removed or subject to  content moderation. This will allow users to understand the reasons that action on  their content was taken, the method used (algorithmic or after human review) and  under which platform rules action was taken. Also they should have processes in place  that permit users to appeal such decisions.  

31. An enabling environment.

This guidance recognises the difficulties of identifying  content that is potentially damaging to democracy and human rights. For example,  separating misinformation) from disinformation (is complex, bringing with it dangers   to free expression of suppressing content legitimately protected under international  human rights law.  

31.1 Platforms should show what they do to provide an enabling environment that  facilitates expression, that challenges false or misleading information, warns of offline  consequences to speech that might be dangerous (e.g., hate speech) or simply flags  different perspectives or opinions. 

31.2 Where possible, users should be given the ability to control the content that is  suggested to them – platforms should consider ways to encourage users’ control over  the selection of the content to be displayed because of a search and/or in news feeds.  The availability to choose between content recommendation systems that display  different sources and different viewpoints around trending topics should be made  available to users in online platforms. 

32. User reporting

In supporting freedom of expression and the availability of accurate  and reliable information in the public sphere, it is critical to empower users of digital  platform services. In this regard, all, companies, governments, civil society  organisations and academic institutions have a roll to play. Companies in particular,  in addition to the platform providing information about its policies accessible in a  digestible format and in all relevant languages, it should show how it allows users to  report potential abuses of the policies, whether that be the unnecessary removal of  content, the presence of violent or threatening content, or of any other content which  is in breach of the policies. Where possible, users should have access to a platform  representative in their own country. 

32.1 The user reporting system should give high priority to content that is  threatening or intimidatory, particularly to groups with protected characteristics,  ensuring a rapid response8One mean could be through an escalation channel for the most egregious threats.   and, if necessary, by providing specific means of filing the  report. This is particularly important when it to comes to gendered online violence  and harassment. A pre-set template would allow the aggregation of similar complaints  that would help identify systemic failings on the platform. At the same time, this  guidance recognises that much of this will depend upon local and regional contexts. 

32.2 There should also be an effective user complaints mechanism to allow users  meaningful opportunities to raise issues of concern. This should include a clear and  understandable reporting channel for complaints and users should be notified about  the result of their appeal. 

32.3 There will clearly be issues of scale for platforms with large numbers of users.  In such circumstances, platforms may need to deploy automated systems to process  and record complaints and the regulatory system will review the operation of these  systems.  

33. Content that potentially damages democracy and human rights, including mis- and  disinformation and hate speech 

33.1 For platforms and independent regulators, attempts to identify potentially  damaging content that is not manifestly illegal can be a significant challenge, as most  freedom of expression legal standards emphasise the importance of context and  intent – saying the same words in different contexts and in different ways can have  very different legal implications. And sometimes apparently legal speech which  constitutes disinformation can be deployed with the intent of causing severe harm.  Different opinions and viewpoints on the potential damage posed by content will  arrive at very different solutions. This is made even more difficult by the sheer volume  of content uploaded continually across all platforms, which can feasibly be managed,  at least in the first instance, mainly automatically. 

33.2 Platforms should say how they define and respond to a wider set of damaging  content through a systematic risk assessment. The regulatory system should assess if  platforms are consistently applying their systems and processes to effectively enforce  their own standards (including the protection of legitimate speech) which should be  aligned to international human rights standards.  

33.3 This guidance recognises the formidable challenge of identifying damaging  speech which may be legal in one context but damaging in another. For example, it is important to distinguish between content promoting hatred directed at women,  children, youth, LGBTTIQ, indigenous groups, people with disability and vulnerable  communities and content that is simply offensive to a particular group of people in a  particular context

33.4 Platforms should demonstrate how they would respond to potentially  damaging speech – either by providing alternative reliable information,9Several media platforms have instituted “disputed news” tags that warn readers and viewers about contentious content   flagging  concerns about the quality of this information, curbing its virality or any other means. 

Content removal or de-platforming of users should be considered only when the  intensity,10The frequency of occurrence over a given period and their range.   and severity11Its scale, scope or irremediability. Scale means the gravity of the impact on the human right(s). Scope means the number of individuals that are or could be affected. Irremediability means the ease or otherwise with which those impacted could be restored to their prior enjoyment of the right(s).   of content that has the intention to harm a group or  individual occurs. The platform should also be explicit about whether it partners with  outside organizations or experts to help it make these kinds of decisions particularly  in countries or regions where the platform itself has little local knowledge. 

33.5 Platforms should show whether they apply specific protection measures to  particular groups. If they do, these measures might include risk assessment and  mitigation processes or the creation of specific products that enable these specific  groups to actively participate online.  

34. Media and information literacy

Platforms should set out the resources they make  available to improve media and information literacy, including digital literacy about  their own products and services, for their users. There should be a specific focus inside  the company on how to improve the digital literacy of its users with thought given to  this in all product development teams. The platform should be reflecting on how any  product or service impacts upon user behaviour and not just on the aim of user  acquisition or engagement.  

34.1 Platforms should implement specific media and information literacy measures for women, children, youth and indigenous groups.12Although this guidance is aiming to highlight the asks of the regulatory systems towards the platforms, it is important to underline that UNESCO has a series of recommendations for governments regarding media and information literacy policies, which must be implemented by relevant authorities, particularly in the education sector.    

35. Election integrity 

35.1 Digital platform servicesshould have a specific risk assessment process for any  election event and should engage with the election’s administrator/regulator (and  relevant civil society groups), if one such exists, prior to and during an election to  establish a means of communication if concerns are raised by the administrator or by  users/voters. Within the assessment, they should review whether political advertising  products, policies, or practices arbitrarily limit the ability of candidates or parties to  deliver their messages.  

35.2 Digital platform services that accept political advertising should ensure in their  terms of service that to accept the advert, the funding and the political entity are  identified by those that place the adverts.  

35.3 The platform should retain these advertisements and all the relevant  information on funding in a publicly accessible library online. Political advertisements  which refer to issues rather than parties or candidates should be scrutinised to ensure  they are consistent with the overarching policies of the platform in relation to hate  speech or speech targeting people with protected characteristics.  

35.4 Digital platform services should adopt transparency measures regarding the  use and impact that the automated tools they use, although not necessarily the  specific codes with which they operate, may have in practice, including the extent to  which such tools affect the data collection, targeted advertising, and the disclosure,  classification, and/or removal of content, especially election-related content. 

36. Major events

Digital platform services should have risk assessments and mitigation  policies in place for “major events” crises such as conflicts, wars, natural disasters,  health emergencies, and sudden world events where mis- or disinformation and hate  speech are likely to increase and where their impact is likely to be rapid and severe. 

37. Language and accessibility

Digital platform services operate globally, and the main  language of many such platforms is English. There are over 7,000 languages spoken in  the world today, though many are spoken only by small groups of people.13https://www.ethnologue.com/guides/how-many-languages. Around 40% of languages are spoken by fewer than 1,000 speakers.   It is critical  if regulation is to be effective that users can operate in a language that they  understand. Setting a reasonable expectation for which languages platforms should  be able to operate in will depend upon the scale, reach, and sensitivity of the service. 

For global platforms, it would be reasonable to suggest that users can contact them  either in one of the six UN languages14Arabic, Chinese, English, French, Russian and Spanish   or in one of the 10 languages spoken by more  than 200 million people.15English, Chinese, Hindi, Spanish, French, Arabic, Bengali, Russian, Portuguese and Urdu.   Automated language translators, while they have their  limitations, can be deployed to increase the number of languages available. Platforms  may wish to ensure the provision of information in additional languages during  election events, perhaps by increasing the capacity of moderation in local languages  during such events. Consideration should also be given to persons with disabilities,  and the ways in which they can interact with, and make complaints in relation to, the  platform.  

37.1 It is recognised that this signifies an important shift in the way platforms  operate, as English is a predominant international language. Nevertheless, as the  major force in global communication, platforms must recognise their responsibility to  allow people to communicate effectively if they are to be accountable to them.  

38. Data access

Platforms should provide stable access, wherever safe and practicable,  to non-personal data and anonymised data. Access should be provided to data that is  aggregated, or manifestly made public data for research purposes through automated  means such as application programming interfaces (APIs) or other open and accessible  technical solutions allowing its analysis. They should provide access to data necessary  to undertake research on content that is potentially damaging to democracy and  human rights and support good faith research that involve their services. There need  to be safeguards with providing data access that ensures the protection of privacy and  respect of commercial confidentiality. For platforms to build reliable interfaces for  data access, there will need to be alignment among regulators that can determine  what is useful, proportionate and reasonable for research and regulatory purposes. 

Section Three – The independent regulatory system  

39. There are vastly different types of bodies involved in online regulation throughout the  world. These range from existing broadcast and media regulators who may be asked  to take on the role of regulating content online, to newly established dedicated  internet content regulators, or general communications regulators given an extended  remit. There may also be overlap in some states with advertising or election bodies,  or with information commissioners or national human rights institutions. Some  regulators may exist independently of government and others might be constituted  as government agencies. It is therefore difficult to set out detailed guidance when  regulation can take so many varying forms and potentially involve so many agencies.  

40. Nevertheless, in whatever form regulation operates, it will constitute what this  guidance calls a “regulatory system” of some kind. The guidance below is therefore  meant to be generally applicable to any system of regulation, however established, 

however varied. Of course, this guidance recognises that this approach, if adopted,  could imply significant changes to the way regulation operates in some Member  States.  

Constitution  

41. Any regulatory system, whether a single body or multiple overlapping bodies, charged  with managing online content (overseeing systems and processes) needs to be  independent and free from economic or political pressures or any external influences.  Its members should be appointed through an independent merit-based process of  appointment, overseen by an oversight body (which could be the legislature or an  independent board/boards). Its members should not seek or take instructions from  any external body, whether public authority or private actor.  

42. The dismissal of members of the regulatory system should be based on clear criteria  and must also be subjected to a thorough process, guaranteeing that their dismissal  is not a result of any political or economic pressures.  

43. The members of the regulatory body/bodies should make public any possible conflict of interest.  

44. The regulatory system must have sufficient funding to carry out its responsibilities  effectively.  

45. The regulatory system should make a regular report to an oversight body on its  findings and will be accountable to it. The regulatory system should also hold periodic  multi-stakeholder consultations on their operation.  

Powers  

46. Regulation will set the overarching goals for platforms to safeguard information as a  public good, empowering and protecting users (particularly vulnerable users such as  children or minorities) and specifying expectations as to how the stated goals should  be fulfilled. It should not make judgements about individual pieces of content but will  focus upon the systems and processes used by the platforms.  

47. While the guidance is developed for those platforms whose services have the largest  size and reach, minimum safety requirements should be applied to all platform service  companies regardless of size.

48. In-scope digital platform services will be required to report regularly on how they are  achieving the goals, though regulators may commission off-cycle reports if there are  exigent circumstances, such as a sudden information crisis (such as that brought about  by the COVID-19 pandemic) or a specific event which creates vulnerabilities (e.g., elections, protests, etc.). 

49. The regulatory system will have the power to call in any digital platform service  deemed not to be complying with its own policies or failing to protect users and, after  discussion, may recommend a specific set of measures to address the identified  failings. Any such judgement should be evidence-based; the platform should have an  opportunity to make representations and/or appeal against a decision of non compliance; and the regulatory system should be required to publish and consult on  enforcement guidelines and follow due process before directing a platform to  implement specific measures. Failing to comply with this stage could lead to penalties  which are proportionate, dissuasive, and effective (but excluding personal criminal  liability).  

50. It will have the power to commission a special investigation or review by an  independent third party if there are serious concerns about the operation or approach  of any platform or an emerging technology. 

51. It is expected that illegal content will be removed solely in the jurisdiction where it is  illegal.16However, it is important to recognise that no systems and processes will be 100% precise in identifying illegal content (at least not without disproportionate intrusion and monitoring). Therefore, it should not automatically be a breach of the regulations if illegal content is found on the service, unless it can be shown that the platform knew of it and failed to report it, or if the relevant systems and processes can be shown to be inadequate. Moreover, identification of illegal content should be interpreted consistently with international human rights law to avoid unjustified restrictions on freedom of expression.    

52. One option as an additional protection for users is for there to be an ombudsman for  complaints about platforms.17The scope of complaints would be limited to a failure to comply with its regulatory duties, rather than as an additional appeal mechanism where users are unhappy with specific decisions.   While in the first instance, complaints should be made  directly to the digital platform service itself, in the event of no or an inadequate  response, the user could go directly to the ombudsman. This may result in an  unmanageable workload, and an alternative for digital platform services with large  volumes of content is for them to have independent complaints/appeals/redress  processes, which the regulatory system can then evaluate. 

53. Given the likely volume of complaints, the regulatory system will be expected to  prioritise those complaints that demonstrate importance and relevance, systemic failings and/or substantial user harm. In this event the relevant regulator will have the  power to intervene and require action, including on an interim/urgent basis if  necessary.16  However, it is important to recognise that no systems and processes will be 100% precise in identifying  illegal content (at least not without disproportionate intrusion and monitoring). Therefore, it should not  automatically be a breach of the regulations if illegal content is found on the service, unless it can be shown  that the platform knew of it and failed to report it, or if the relevant systems and processes can be shown to  be inadequate. Moreover, identification of illegal content should be interpreted consistently with  international human rights law to avoid unjustified restrictions on freedom of expression. The scope of complaints would be limited to a failure to comply with its regulatory duties, rather than as an  additional appeal mechanism where users are unhappy with specific decisions.

Review of the regulatory system. 

54. There will be provision for a periodic independent review of the regulatory system,  conducted by a respected third-party reporting directly to the legislature and  subsequent consideration by the legislature. 

55. Any part of the regulatory system should act only within the law in respect of these  powers, respecting fundamental human rights – including the rights to privacy and to freedom of expression. It will be subject to review in the courts if it were believed it  had exceeded its powers or acted in a biased, irrational or disproportionate way.  

56. Decisions on eventual limitations of specific content should be taken by an  independent judicial system, following a due process of law.