UN Special Rapporteur on Minority Issues [DRAFT FOR GLOBAL CONSULTATION] The ideal standard adopted by SMCs and required by States should ideally go higher as stated above in line with the aspect of online speech to proliferate at speed, the extent of the potential response as removal of community members from one platform. This would in practice mean that the threshold of criminal law laid out in RPoA should not need to be met for incitement to be applicable in the context of SMC content moderation. This has a number of facets. Firstly, all 6 parts of the test need not be satisfied for incitement to be found. As such if focus is placed on protecting protected groups and especially minorities from hate speech or incitement to hatred, emphasis should be placed on the outcome. Intent may be replaced with ‘effect’ or ‘likelihood’ and ‘imminence’ with ‘some causality and risk’ (UNGP). It should also be noted that fulfilling some of the 6 parts in the online context need not be as difficult to satisfy as in other ones. These include the idea of ‘reach’ and ‘status’ in the online setting need not attach to a prominent personality, figure or politician with a vast dedicated following. Such expressions will now doubt be even further amplified, but it does not discount the possibility of a publicly unknown individual gaining an audience of millions to acutely hateful online rhetoric. The social media context also necessitates the introduction of a further aspect not mentioned in RPoA and relevant to the UNGP of exposure which is a product of virility multiplied by time. This underscores the vast reach and instantaneous speed at which hate and incitement to hatred can proliferate. A separate dimension to the types of expression which may be included in incitement policy is the necessity to expand the definition to include incitement to discrimination and hostility, not just violence. Such a distinction is arbitrary and becomes engaged too late in the process of escalation envisaged by ICCPR Art. 20(2) in that what culminates in violence begins with incitements to discriminate and then to hostility. Therefore, it remains imperative for an approach focused on the protection of protected groups and minorities from incitement to hatred that results in eventual violence that protection should also be offered beginning that the level of incitement to discrimination. In aspiring towards this higher level of protection of human rights that SMCs can seek to implement it could be argued that any statements which are discriminatory in nature or may normalised discriminatory attitudes towards protected groups and minorities should be limited. It must be emphasised that any limitations on online expressions based on an ‘incitement’ policy must satisfy the three-part test for it to be permissible in compliance with the right to freedom of expression legality, legitimacy, and necessity and proportionality (elaborated and contextualised in detail above). Ultimately, any such ‘incitement’ content policy must ensure that whatever criteria are employed, the protection offered does not become illusory to such an extent that it never practically succeeds in preventing violence, but rather only becomes engaged as an enforcement response once the violence has occurred. This is the ultimate test of an effective ‘incitement’ content policy. 5. SMC’s should publicly make available how they identify content that falls short of hate speech content policies and is algorithmically demoted, and likewise steps in place to ensure hate speech is not promoted or becomes virulent prior to removal. Commentary SMCs are responsible not just for the presence of user-generated content on the platforms but how that content is curated, packaged, presented, promoted and amplified. In this sense, the harm emanating from ‘hate speech’ and ‘incitement to discrimination’ targeting minorities does not just depend on its severity but also its reach, which in turn is determined by a number of automated systems employing artificial systems reliant on complex algorithms. Historically, the greatest criticism of SMCs has been that their design was intended to cause controversy, adversarial interactions and create self-reinforcing and polarised echo chambers. This had positive ramifications through the stimulation of democratic public debate and for community members to form groups of likeminded members with similar 11