The General Scheme of the Online Safety and Media Regulation Bill

Point 5 states that the Media Commission will “provide a regulatory framework that takes account of the rapidly changing technological environment and that provides for rules to be applied in a proportionate, consistent and fair manner across all services regulated, having regard to the differing nature of those services.” As far as possible, new media legislation should be future-proofed to accommodate technological change and developments in the media sector. This may be achieved through a focus on outcomes and objectives (e.g. online safety, media plurality) rather than prescriptive obligations; through a medium-neutral approach to technology and media; and through built-in review mechanisms.

Point 2(a) refers to “linguistic, religious, ethical and cultural diversity”. We presume ‘ethical’ should be ‘ethnic’.

This head outlines the overall functions of the Media Commission. It states that “the delegation of functions is ultimately a matter for the Commission itself” and that “individual Commissioners can take responsibility for clearly delegated functions [which] is particularly relevant in the case of the Online Safety Commissioner”. However, as the Media Commission will have wide-ranging powers that greatly exceed those of the Broadcasting Authority of Ireland, greater clarity about the commissioners and their roles and functions is desirable. Specific responsibilities are discussed in the appropriate heads below.

Given the increasing (if belated) acknowledgement within political circles of the constitutive role played by media outlets in the operation of liberal democracy in Ireland, some consideration should be given to how the Media Commission might overtly seek to support this media function. We suggest that, in addition to the Online Safety Commissioner, the role of Media Pluralism Commissioner be introduced. The role of this Commissioner would be to consider how the policy and regulatory environment in which all Irish-facing media operate can be best designed to create and maintain a healthy, pluralistic and diverse public sphere. We use "pluralism" here in its broadest sense: i.e. we are not merely concerned with questions of media ownership and the concentration of same. The Media Commissioner would be concerned with:

Much of this already falls within the remit of the existing Broadcasting Authority of Ireland. Furthermore, some areas are - rightly - the primary responsibility of other regulatory bodies (such as the Freedom of Information Commissioner). Nonetheless, we suggest there is a logic to grouping these policy areas together as they cumulatively constitute a coherent basis for media pluralism understood in its broadest sense. Thus, even if the Media Pluralism Commissioner did not have direct regulatory responsibility for all of these areas, we would anticipate that the Commissioner would be actively factored into deliberations around the formation of any new sets of regulations.

Section 1 provides that “the Commission, in the interests of the effective discharge of its functions, may enter into cooperation agreements with other bodies as it sees fit”. Given the scale of the challenge and the complexities involved, we recommend that certain bodies - vetted experts or regulatory bodies from other Member States - be given the status of “priority complainant”. This status would create a responsibility to pay special attention to the considerations raised by these complainants.

Disinformation has been excluded from the categories of harmful online content. We recommend this Bill should include a Head on Disinformation to ensure that there is a specific responsibility to tackle harmful disinformation. Omitting disinformation undermines the Media Commission's objective to "ensure that appropriate regulatory arrangements and systems are in place to address where appropriate, illegal and harmful online, sound and audio-visual content". 

This omission will have significant implications for individuals, communities, and public safety and, contrary to EU obligations, it will leave this area unregulated in Ireland.

Disinformation is defined as the deliberate creation and/or dissemination of false or misleading information for personal, political, or financial gain¹. Although disinformation is not illegal, it can have harmful impacts on individuals and society. These harms were evident throughout the Covid-19 pandemic as online platforms provided a breeding ground for false and misleading information with significant implications for individuals, communities, and public safety².

In response, the European Union’s Democracy Action Plan and Digital Services Act call for coordinated action among Member States. In 2021, the European Commission will revise the 2018 Code of Practice on Disinformation³, which placed reporting responsibilities on digital platforms regarding the transparency of political and issue-based advertising and a commitment to address fake profiles and accounts and the demonetisation of disinformation. For example, the FuJo Institute was commissioned by the Broadcasting Authority of Ireland to write two reports - Elect Check⁴ and Code Check⁵ - on the implementation of the Code in Ireland. These reports highlighted important inconsistencies across digital platforms.

Beyond the Code, the European Union calls for increased investment in media literacy as a means to build societal resilience to harmful disinformation. Actions in these areas overlap with the proposed functions of the Media Commission; specifically in relation to the impact of automated decision-making by online services; media advertising, the promotion of media literacy; the assessment of harms; and the development of media codes. As such, the Media Commission is ideally placed to address the issue of harmful disinformation in cooperation with EU Member States. In contrast, the proposed Electoral Commission Bill is restricted to a narrow focus on misleading advertising during election campaigns.

Heading b provides for an example of harmful content that appears to pertain to harassment and it is close in its description to cyberbullying: “(b) material which is likely to have the effect of intimidating, threatening, humiliating or persecuting a person to which it pertains and which a reasonable person would conclude was the intention of its dissemination.” The explanatory note further clarifies: “The other included categories of material refer to cyberbullying material pertaining to any person, which includes all kinds of cyberbullying material, material promoting self-harm and suicide and material promoting eating disorders” and “the second category of material included under harmful online content is intended to encapsulate the notion of cyberbullying. This category has a base in Article 28b(1)(a) and (b) of the revised Directive. Subparagraph (a) concerns material which may ‘impair the physical, mental or moral development of minors’ and subparagraph (b) concerns the ‘incitement to violence or hatred… based on any grounds referred to in Article 21 of the Charter.’”

It might be helpful to keep in mind the definitions of cyberbullying among children when deciding on the specific examples and categories of such content a) in order to ensure that cases relevant to the experience of children and teens are covered; b) in order to properly incentivise the platforms to provide the right assistance to children and teens when they are in the position to do so. The lines between harassment and cyberbullying can be blurry and the ways in which these cases can manifest on a variety of platforms are diverse. This is why, bearing in mind the following definition of cyberbullying might be helpful when thinking specifically about the protection of children across online services and platforms and whether your proposed example of intimidating, threatening and humiliating behaviour captures the entirety of the issue. Cyberbullying definitions tend to be derived from definitions of face-to-face and school bullying.

School bullying is defined as a form of aggressive behaviour which aims to harm a chosen target⁶. Although various definitions of bullying can be found in the literature, scholars tend to agree on three main features of bullying: 1) Intentionality: Bullying is a goal-oriented and systematic abusive behaviour; 2) Repetition: Bullying happens repeatedly and 3) Imbalance of power: The systematic abuse of power is perpetrated by someone who is either physically or psychologically stronger than the target. Bullying has a negative impact on the target, in terms of their psychological and physical wellbeing. Offline bullying can take physical, verbal and relational forms (e.g., social exclusion, gossiping) and the latter two are manifested variously in online environments. Offline and online bullying are highly correlated; i.e., cyberbullying is an extension of traditional bullying as targets often face the same perpetrators offline and online⁷.

It is our understanding from the text of the OSMR bill and from the meetings of the National Advisory Council for Online Safety (NACOS), that this piece of legislation intends to address what has elsewhere been termed as “serious online abuse and harassment.”⁸ If your intention is to regulate serious cases of harassment and cyberbullying ---there needs to be a clear understanding as to what “a serious case” actually is (even though the word “serious” is not in the text of the Bill). Several subtly negative comments that are perhaps ironic in character might not fall under the definition proposed in the OSMR Bill, yet they could have a serious negative effect on a child or a teen; and they can also be part of a broader ecosystem of cyberbullying happening on several platforms simultaneously; and such comments can be just one component of a larger offline incident. Therefore, it is essential that the issue be addressed not only at the level of content takedown; but also at the level of content distribution on many platforms. This means that working in a coordinated way to prevent bullying content from spreading from one platform to another should be considered; and also to consider the proposition that content that is labelled as bullying or harassment should receive less visibility and priority in the algorithmic curation of content on a single platform.⁹ Social media platforms are developing artificial intelligence-based tools that are able to detect these subtle comments; and they can provide tools for children and teens to not only avoid exposure to such content; or leverage help from peers/friends, parents and educators; but also to downplay the visibility of such content and affect its circulation.¹⁰ Such actions should fall under online service providers’ duty of care, as long as they do not constitute an obligation to monitor content, which is prohibited under eCommerce directive and subsequently DSA.

With regard to Section 8 point d (“the impact that the nature and prevalence of certain harmful material available online may have on users of relevant online services and the general public”) and point f (“the impact of automated decision making in relation to [content delivery and content moderation] by relevant online services,” we would like to underscore that there is insufficient research as of now addressing these points when it comes to minors and cyberbullying¹¹; and it is important to solicit the collaboration of social media platforms in either providing anonymised data to examine these issues or to provide evaluation of their automated decision making processes and moderation. Some social media platforms are large entities where a number of different teams might be working on different aspects of one issue and they may not always exhibit sufficient coordination and even awareness of each others’ work¹². In a recent attempt to conduct research interviews with representatives from a number of social media platforms about artificial intelligence-based proactive moderation (i.e. detecting cyberbullying content before it has been reported by a user who is a child), one of the authors of this submission was not able to receive responses from a number of large social media platforms. Some companies, such as Facebook and Twitter, list the percentage of cyberbullying cases that are detected proactively in their Transparency Reports.¹³ In order to ensure transparency leading to accountability, we find it important that companies not only disclose the percentage of cases that had proactively been taken down; as such numbers might not be meaningful without more context (e.g. is 30% of cyberbullying cases taken down proactively satisfactory; were children who were bullied sufficiently assisted by such proactive content take-down; what if the content had already spread to other platforms; what are the implications of such takedowns for freedom of expression).

Companies need to be able to explain how their AI models that drive these takedowns are developed and enforced for at least three reasons: 1. We need to be able to assess their effectiveness independently from what the companies are stating/publishing 2. We need to examine their effectives from the perspective of children 3. We need to be able to understand freedom of expression and privacy implications of such proactive monitoring. This is why we find it important that the government mandates evidence and explanations regarding these issues via investigations and audits.

We note that these issues also apply to other areas of harmful content including disinformation where there is a lack of transparency and accountability regarding the role of algorithms in content moderation decisions and a lack of oversight regarding the potential for unjustified content or account removals.

Using age as a regulatory and organizing instrument to signal maturity is seen as an effective way for adults to manage children’s access to technology. Nonetheless, such systems do not allow for a sufficient level of nuance in terms of individual differences. Age-gating considers all children in one age cohort to be the same while children’s maturity levels within one age cohort may differ. For example, thirteen years of age has long been the cut off point for social media use, decided upon as the digital age of consent for privacy purposes during the making of the Children’s Online Privacy Protection Act (COPPA) in the United States in 1998; and this age limit is not based on extensive research, it merely signals the onset of adolescence. When enforcing age-gating as a measure to prevent exposure to age-inappropriate content, it is important to consider the balance of children’s rights to protection on the one hand and provision and participation on the other. Encouraging some level of content-labelling done proactively by platforms (e.g. covering potentially harmful content so that users can choose whether to see it; encouraging or even requiring users to age-restrict their videos as in the case of YouTube) might be an option to consider as well. Children’s levels of maturity vary, and the enforcement of age-based content restrictions can pose challenges for the balance of rights, most recently seen in the debate around the application of the GDPR Article 8¹⁴ (Macenaite & Kosta, 2017). In light of the recent General Comment on the application of the United Nations Convention on the Rights of the Child (UNCRC) in digital environments,¹⁵ it might also be advisable to implement a Child Rights Impact Assessment (CRIA) into the Online safety codes and compliance assessments.¹⁶ Furthermore, some parents and guardians might think that the age of digital consent conveys the message of when it is safe for a child to use a platform i.e. that the age of 13 (i.e. 16 in Ireland) is a safety advisory recommendation rather than it being related to data and privacy protection. This is why it might be helpful to encourage education of parents, guardians and educators with respect to the meaning of age-gating and digital consent. Furthermore, the enforcement of COPPA and Article 8 of the GDPR has thus far been easy to bypass and it is well known that there are a number of underage users on social media platforms that the companies do not recognize and take into account17. Addressing the issue of underage use and encouraging platforms to innovate for underage users in a regulatory environment that incentivises platforms’ denial of their existence on the platforms, remains a challenge.

When creating online safety codes, we find it particularly important to focus on point 3, sections d (the nature and scale of designated online services or categories thereof); e (necessity for transparency of decision making with respect to content delivery and content moderation) f (the impact of automated decision making), and g (the nature and prevalence of harmful online content) in the context of the protection of minors from harmful online content.

Section 4: “the Media Commission may examine the compliance of designated online services with online safety codes on the basis of the information requests specified in subsection (1) and
other information that Commission considers relevant, including matters brought to the
attention of the Commission by a nominated body under Head 53B or other interested
parties such as members of the European Regulators Group for Audiovisual Media Services.”

If applicable: We recommend that interested parties include not only nominated bodies but also academic institutions and independent research organisations, think tanks and non-governmental organisations (NGOs) or charities, if these are not encompassed by this provision already.

Having in mind the recent decision on the application of the UNCRC in digital environments, if it is suitable, we suggest that a point be added to section 3 which says that in preparing these guidelines with respect to protection of minors, the Commission should take into account the balance of children’s rights to protection, provision, participation and privacy; furthermore that children be consulted through research for the purposes of creating these guidelines materials?

If possible/needed at all, we suggest to add specifically a provision that bodies such as academic institutions, researchers etc. might provide advice/assistance/evidence in auditing complaints handling.

While the current version of the Bill does not provide for an individual complaints scheme - present in the Australian Online Safety Act - due to concerns that the transposition of the AVMSD would place Ireland in the position of responsibility for complaints spanning many EU countries; we are of the opinion that the government should reconsider this approach in order to ensure that there is a “mechanism to vindicate the [AVMSD] Article 28 (b)(3)(i) procedures regulating reports/complaints made by minors to the platform providers.” This position is supported by the Irish Society for the Prevention of Cruelty to Children (ISPCC) which obtained a legal opinion on the matter that we are quoting from here.

Additionally, as highlighted in our remarks around transparency of decision making under Head 50A, we are concerned that the lack of an individual complaints scheme may lead to an underreporting of issues in cases where individuals are unhappy with the results obtained from a provider’s reporting process.

We find this section to be a bit unclear. We would like to see some clarification regarding what type of case might require meditation and under what rationale might the costs of mediation be borne by the end user? Additionally, we are concerned that in the absence of an individual complaints mechanism, end-users who are dissatisfied with a platform provider’s response may be deterred from pursuing this further if the option provided to them is a complicated and potentially costly mediation process.

We would like to ensure that there is no definitional confusion about the term “media service provider” and the competition law definition of a “media business” in the 2002 Competition Act. The latter definition would imply that the proposed regulatory provisions only apply to media that have sales in excess of €2 million in the State in the most recent financial year. However, under some readings, the current text could imply a much wider definition of "media service providers" that includes all those providing AV or sound media services such as, for example, individual podcasters.

It is worth noting that the reach of the levy may not be quite as extensive as envisaged given the impact of Brexit. Since the legal basis for the content levy is based on Article 13 of the revised AVMS directive, its application is limited to audio-visual media service providers based within the EU. On linear television there are currently at least 36 channels offering advertising slots targeting the Irish market. These obviously include Irish-based channels such as RTÉ and Virgin Media. However, the bulk of these 36 channels (including Channel 4 and Sky) are based in the UK and thus fall outside the regulatory reach of the new Media Commission and the AVMS Directive

In consequence, in considering the viability of a content levy, the focus of the new Media Commission will also certainly be on streaming services. According to the Mavise Database there are over 223 licenced audiovisual media services based in Ireland. The majority of these are online audiovisual service providers targeting other EU Member States: Apple TV+, Facebook Watch, Google Play, Instagram, the iTunes Store, the Microsoft Store, MSN, and the various incarnations of Youtube. All of the above also offer services targeting Ireland and these along with Netflix (based in the Netherlands) would be potentially subject to the levy.

However, any research into the viability of a levy would have to consider how to disaggregate the revenues earned by, for example Apple TV+, across the 27-country-specific operations which are licenced to operate from Ireland. A further consideration might be whether the introduction of the levy would create an incentive for such operations to relocate beyond the borders of the EU so as to evade its reach.

However, any research into the viability of a levy would have to consider how to disaggregate the revenues earned by, for example Apple TV+, across the 27-country-specific operations which are licenced to operate from Ireland. A further consideration might be whether the introduction of the levy would create an incentive for such operations to relocate beyond the borders of the EU so as to evade its reach.

There are two macro considerations that should be taken into account here: (1) the logic of AVMS Article 13 and (2) parallel discussions under the auspices of the Future of Media Commission.

As currently stated, the Content Levy Scheme essentially cuts and pastes from part 10 (“The Broadcasting Fund”) of the 2009 Broadcasting Act. This places limits on the kind of material that might be supported by the content levy. It completely excludes news and current affairs content and places a strong emphasis on material with a strong connection to “Irish culture, heritage and experience”. The rationale for this narrow cultural focus is not clear, especially given that - in contrast to the existing Sound and Vision scheme - the fund is being sourced from exclusively commercial (as opposed to publicly-funded) sources. There is nothing in Article 13 of AVMS which even hints at the need for such strictures and they may hamstring the capacity of Irish production companies to create content that is attractive to streamers seeking content which will permit them to meet 30% European content quotas.

There is a risk that streamers will seek to meet such content obligations by adopting the “quota quickie” approach that characterised British film production from the 1930s where distribution and exhibition quotas drove the production of low-budget - and often concomitantly low-quality films - purely to allow distributors and cinemas to meet quota obligations. Content narrowly themed around Irish culture may allow streamers to meet quantitative quota targets across the EU but there will be little incentive to invest substantial resources in such content unless it also promises to meet the commercial objectives of streamers.

In this regard it is worth noting that pay-per-view services like Google Play and subscription services like Netflix and Amazon Prime have different business models. These business models may have a different impact on their demand for European content under Article 13 of the AVMS. The subscription model attracts audiences on the basis of bundling a variety of content. The trick is to keep the bundle sufficiently varied so as to limit the incentive for audience “churn” (i.e. the possibility that subscribers will not renew a subscription at the end of a subscription period). Eclectic Irish-themed content might conceivably constitute an element of this, but it is clearly less likely to alienate European audiences if it is less culturally specific. As regards pay-per-view services, it is hard to see what incentive they might have for including Irish-focused content in catalogues offered to non-Irish audiences (beyond meeting obligations to meet content quotas). Moreover, it is worth recalling the broader objectives of the current “Audiovisual Action Plan” to establish Ireland as a global hub for audiovisual production. The narrow focus on the kind of Public Service-oriented content referred to in Head 77 does not appear to fit very well with this ambition.

However, if there is to be a focus on Irish content, it may also be worth considering whether the content levy might be used to fund support for medium-neutral public-service content production as has been proposed in several submissions to the Future of Media Commission. This might conceivably see such funding used to support content production across ALL media: print, radio and television (linear or on-demand) and online. It is not absolutely clear that Section 13 requires that funds raised through the content levy necessarily have to be spent on audiovisual production. Section 13 (2) states: “Where Member States require media service providers under their jurisdiction to contribute financially to the production of European works, including via direct investment in content and contribution to national funds, they may also require media service providers targeting audiences in their territories, but established in other Member States to make such financial contributions”. This does not appear to exclude support for other media content production. Finally, in light of the submissions to the Future of Media Commission, a wider vision for funding in support of the Irish media sector, including journalism, might look beyond funding content to support infrastructural and resource needs through provisions for technology investment and media training.