This commentary on the European Commission’s Communication ‘Tackling Illegal Content Online’ has just entered my top 10 all time posts. At 8,500 words admittedly it’s on the long side. So here goes a threaded summary. cyberleagle.com/2017/10/toward…
1/20 The EU Council Freedom of Expression Guidelines stress the importance of “protecting intermediaries from the obligation of blocking Internet content without prior due process.”
2/20 ‘Tackling Illegal Content Online’ institutionalises the opposite: prior restraint instead of prior due process. Small wonder they deleted a previously leaked draft’s reference to the Guidelines.
3/20 According to the Communication, platforms should act as gatekeepers rather than gateways: proactively filtering users’ posts and in some cases removing content on the say-so of ‘trusted flaggers’.
4/20 But even law enforcement bodies are not legally competent to make illegality decisions. Police are not the courts. A trusted flagger’s subject matter familiarity is not a substitute for due process.
5/20 Instead of due process at source the Communication advocates quality assurance, internal training and process standards for trusted flaggers, as if assessing the legality of what we write online were like grading peas in a processing plant.
6/20 Assessing most illegality requires knowledge of context. Awareness of content is not the same as knowledge of illegality. Unsurprisingly the Communication struggles badly with the problem of context.
7/20 Hate speech and terrorist material are among the most highly emphasised targets of the Commission’s scheme, yet are inherently vague offences.
8/20 Where speech is concerned even the clearest laws tend to fuzziness. No amount of training – of computers or people - can turn qualitative evaluation, about which humans may disagree, into a precise scientific measurement.
9/20 A system that asks only “legal or illegal?” does not take account of whether a breach is sufficiently serious to warrant removal.
10/20 Nor does that binary question take account of broad criminality definitions intended to be mitigated by prosecutorial discretion.
11/20 Due process at source guards against overreach and institutional groupthink. From a fundamental rights perspective reversibility safeguards may not sufficiently compensate.
12/20 The Electronic Commerce Directive is a liability shield not, as the Communication suggests, a general legal framework for illegal content removal.
13/20 Article 15 ECD prohibits a Member State from imposing a general monitoring obligation on a host to seek out illegal content or activities.
14/20 If a host does not remove content expeditiously when it becomes aware of unlawful content or behaviour, it loses the shield. The host may or may not then be liable under a Member State’s substantive law.
15/20 The host remains free to 'publish and be damned'. The ECD places no positive duty on the host to take action. The Communication aims to reframe the ECD as a removal regime in which platforms are positively expected to act as arbiters.
16/20 These ECD rules, even if they risk incentivising cautious over-removal by hosts, embody a degree of respect for due process and the presumption against prior restraint.
17/20 The Communication makes no attempt to resolve the tension between a coherent EU-wide voluntary removal process and the significantly varying substantive content laws of EU Member States.
18/20 The mantra that what is illegal offline is illegal online does not support removal mechanisms that would be rejected in the offline world.
19/20 Filtering and automated takedown process requirements introduce a level of prior restraint online that does not exist offline.
20/20 If the argument is that the internet is different and requires stronger ways of controlling or suppressing speech, that should be made overtly rather than behind the banner of offline equivalence.
