Trump against Twitter He has a Point. And this is a challenge
by Detlef Eckert
2 July 2020
The story about Twitter that dared to label tweets of Donald Trump should be well known by now. At the time of writing: Facebook disagrees with Twitter by stressing freedom of speech. As a result of advertisement boycott of big companies, Facebook announced it would change its approach. Twitter says it applies different rules to politicians than to ‘ordinary’ people, but has now announced it would block certain posts, for instance, misinformation about voting or racial discrimination. Snapchat stops promoting Trump’s account on its discovery platform. The online platform Reddit has closed 2,000 chat groups/communities, where people exchanged extremely radical and dangerous views, amongst them a group called The_Donald, which has nearly 800,000 users. The company says: “Reddit is a place for creating community and belonging, not for attacking marginalized or vulnerable groups of people.”
“Racial violence and injustice have no place in our society and we stand together with all who seek peace, love, equality, and justice in America,” Snapchat.
Trump argues that platforms are neutral and therefore should refrain from interfering with ‘his’ posts or should otherwise face regulation as media. The schizophrenic logic behind this argument coming from Trump is obvious. His anger of being ‘assessed’ by a social media platform brings back the question of what role these online services should have in our society. It laid open issues of free speech, the manipulative power of social media, and the regulatory responses. I am entering this discussion for historical reasons. We should see this debate from the perspectives we had at the beginning of the internet boom. At the end of 90s we—policy advisors in the US (i.e. the Clinton/Gore administration promoting the information superhighway) and in the EU (i.e. the eEurope initiative as part of the Lisbon agenda) - were promoting the internet. The US and the EU led the world inter alia towards liberalisation of telecommunications, removed barriers of the use and export of encryption technologies, and promoted e-commerce through a ‘platform regulation’ which exempted e-commerce and internet service providers from liability of content hosted or passed through their networks. Yes, one of this regulation is the famous e-commerce Directive, which the EU will review soon. All of this worked well at the beginning.
The positive impact of the internet on media, access to information, and democracy thrilled us. We believed nobody could control the internet and free speech would conquer the world. Autocratic governments could not stop this free flow of information. Finally, citizens were no longer bound to traditional media and their political affiliation. The ‘Arab Spring’ seemed to confirm this ‘freedom bringing’ effect of the web and social media. Soon we were to find out a distinct reality.
Radical Islamic groups show how effectively one can use social media platforms to distribute evil propaganda and brain-wash people. Citizens—instead of searching for the truth—stay in their bubbles, looking for confirmation of their prejudices. Software bots are powerful instruments to distribute disinformation at large scale. We experience hate speech and attacks on people with a different opinion. No, we did not predict this habit called ‘shitstorm’. China proves that it is possible to filter content going in and out of the country. How likely would be a President Trump without Twitter and Facebook? Artificial Intelligence will enhance these manipulative activities.
The notion of neutral platforms has never been accurate. ‘Notice and take down’ regulation, in particular to satisfy the need of the right holders, obliges service providers to act on content. Law makers across the globe have issued rules which kind of content is illegal to post. However, here we must distinguish between governments taking action against their citizens as breaking the law and the platforms. The problem is the global reach of platforms. Posting Nazi-propaganda may be illegal in Germany but not in the US, whilst certain erotic material is fine in Germany but not in the US. The Internet Government Forum discussed at length what makes up illegal and harmful content, and so far, the only topic on which everyone agrees is ‘child pornography’.
Even abstracting from regulation, platforms are not neutral entities. They have established their own policy and rules. Within the boundaries of applicable law they decide what is permissible to be posted and what not. Users have to accept or move to another platform. And here we are getting closer to the point behind the skirmish between Trump and Twitter. Network effects create dominant platforms. The more users a platform has, the more value it provides for each user. Google+ never took off, Facebook had already too many users. As users have de facto less and less choice, freedom of speech and permissible content is in the hand of private companies, in fact monopolies. Most of these globally dominant platforms are US-based. The only country that has managed to set up its own platform eco-system is China. Because of its power, it can apply censorship.
Coming back to Trump. Twitter allowed politicians full freedom of speech unlike with ordinary citizens. It argued that it was in the public interest to know what politicians think, even it was against the company rules. With tagging Trump’s posts, it introduced a new policy, which is a tough line to draw. What qualifies to add a qualifier? In addition, fact checks are not necessarily neutral. I am not talking about ‘alternative facts’. I refer to the experience that sometimes empirical evidence differ.
What about Europe?
The EU is preparing a Digital Service Act, with which they will put a regulatory framework around platforms. European politicians think requiring these platforms to control, and yes to filter the content of their users, even with upload filters, is not censorship but enforcing the law and protecting the democracy. For instance, France issued a ‘hate speech’ law to do exactly this. However, the French Constitutional Court ruled this law as unconstitutional.
This shows the dilemma, when implementing rules concerning online content. First, to differ from ‘notice and take down’ rules, legislation must oblige platforms to filter content without explicit notice from an authority. Think about its implementation. For instance, what if a platform could not identify false information or bots disguised as people? What do we want them to do to distinguish disinformation from a foreign power from a citizen with extreme political views? To be effective, regulation has to enter deep into the platform business with a strong oversight. My conclusion is that the EU need to establish a ‘forced’ partnership with dominant platforms by implementing a close ‘notice and take down’ system. Leaving the platforms alone to implement complex and context-dependent rules will not be effective and as the French case shows can become ‘unconstitutional’.
Second, how to address the global implications of platform regulation? We will face not only different but incompatible rules between regions and countries. Platforms will look differently depending on the country from where you access them. This is already the case with authoritarian regimes where access to platforms or unwarranted content is blocked. It will soon become reality with the rest of the world. The idea of late 1990s internet enthusiasts that the Web would open the world to free speech and better informed people is becoming quickly a faint memory.
What are platforms? Media or neutral repositories and distributers of content? This is the point Trump implicitly put on the table. The answer is: they are none of them. They have never been neutral, and they will never be media as they do not curate content. Regulation is a two-edge sword, necessary to contain the worst effects of platforms, but introducing the bitter pill of ‘content-control’, even we do not want to call it that way.
What to regulate? The consultation and the impact assessment will give more clarity. Some of my reflections: Why not requesting more identification before allowing users to post? Do we need anonymity to express our views? We need more transparency regarding sponsored political advertising. Platforms should use AI to find and disable botnets. Regulation could also make some fact checking mandatory. Platforms should flag AI generated or manipulated texts, images and videos.
So, I am back where it started, namely the dream of a global internet and open access to content. Platforms connecting people is a good thing, but as gatekeeper to information and creator of opinion bubbles they are damaging. Regulation may put a lid on some platforms, but people that like to join extremists group, will move to another one. The members of The_Donald announced they would ‘meet’ at theDonald.win.
The internet provides access to millions of Websites, including those from traditional media. People must learn how to use them and avoid relying on dominant platforms as the only source of information. Therefore, the ultimate answer to hate speech and damaging content is education. We need to introduce digital competence in school curricula and adult learning to fight the challenge of dominant platforms, disinformation, and freedom of speech.