Contrary to theoretical promises of the Internet as a technology to diversity information sources, the lived experience of the majority of the population is anything but diverse. Despite the immense opportunities for variety that the technologies powering the Internet present, two thirds of individuals who consume news on the Internet rely on three main platforms; Facebook, YouTube, and Twitter.

This is not unique to the Internet, as the histories of print, radio, and television illustrate. From the many options that emerge after a major technological breakthrough, only a few last in the competition for limited audience. Complex mechanisms have conspired to select just a few of these platforms as the Internet-experience of our age, creating platform dependency. While they are aggregators of content from other publishers, their ‘gatekeeping’ role in shaping what people have access to has grown significantly, thanks to their business models of paid-for hierarchical placement of content, and their adjudication of source trustworthiness based on original language, source jurisdiction, among others attributes.

The dangers of centralised Internet gatekeeping have been all too apparent to critical observers, but it took the November 2016 elections in the United States of America to awaken popular attention. The tentative charge seems generally pointed to the disproportionate influence Internet platforms have, combined with their laissez faire attitude towards content manipulation. To that end, the proposed solution to dealing with disinformation in a centralised Internet is enhancing content moderation by platforms.

But if this disproportional centralisation is strong enough to significantly alter information flow in countries with historically strong institutional practices, platform dependency takes radically different forms in emerging democracies, presenting gaping vulnerabilities to democratic processes during and in-between elections. As a necessary step towards understanding disinformation in developing countries, the policy conversation has to expand from regulating the narrow profit-minded platforms, to include the role of the most important source of public information in these societies, the State.

The State, through its institutions, is involved in almost all aspects of public information cycles. Public decisions, establishment of social facts, and custody of consensus during uncertainty are core domains of State institutions. Elections, security and law enforcement, monetary policies, disease outbreaks, foreign policy developments, and population statistics are but some of the areas that involve the State at their every turn.

Government communication therefore becomes a core pillar of a reasonably efficient society. An institutionalised government communication process, provides for an informed citizenry while a poor one leads to vacuums and uncertainty. Uncertainly on public information, at least in principle, is more pronounced in developing countries than in established ones. As Internet usage and the popularity of social networks grew, government communication has increasingly moved from state-owned assets to digital accounts hosted on Internet platforms.

The United Kingdom created Government Digital Service, (GDS), as a continuation of its information office. It took up the online experience with institutional memory, expertise, and assets that come with a government office. US.GOV in the United States is also the official extension of Government Information and Services. This is the case with Canada, the EU, and refreshingly with Rwanda, among other countries. In contrast, most governments I have studied in the African region have very active Facebook pages, some Twitter accounts, yet these are not mirrored on their own websites. Social media platforms in this region are not necessarily amplifying government communication infrastructure. They are replacing it.

Platform Dependency Deeply Coded

Even while the platforms offer opportunities for innovative government communications, at times being the only window into government, they raise further problems. Some of these online accounts reporting on government developments are under personal control. It gets hard to separate the official from the personal. Archiving of such information is problematic, and freedom of information requests, for example, may never get to vital government information on virtue of being lost between personal and official accounts. On a technical level, securing such highly-prized distributed accounts gets more complicated.

After the current Ethiopian Prime Minister took over leadership in April 2018, Fitsum Arega, his close confidant, used his personal twitter account to communicate Office of the Prime Minister (OPM) developments in his role as the Chief of Staff. Journalists, diaspora communities, and diplomats hang on to his tweets, in large part due to limited timely official alternatives. After fighting numerous troll accounts, including flagging them for impersonating the Prime Minister, he had to apply to Twitter for verification. It took several weeks for the Chief of Staff of Ethiopia to be verified by a platform set up a decade ago. It is important to note that Twitter does not have any office in Africa, and the whole process felt more like a begging instance. Despite all the setbacks, he proved to be a reliable source of information during the crucial transition period. In November 2018, Fitusm was transferred from the OPM. He has hypothetically left with his 50,000+ Twitter followers. Institutional memory, gone. The new press secretary and her team will start that complex digital journey afresh, albeit with occasional retweet boosts from the now verified Fitsum account.

This Ethiopian scenario is not an isolated case, as most governments in the region struggle to manage communications on the Internet with a semi-broken or non-existent government digital infrastructure. Even in countries with a seemingly organized digital communication strategies in the region, they are more or less the president’s account- the person - broadcasting unit, funded from public pulse, with their partisanship evident during election periods. Rwanda, Kenya, Cameroon, and Burundi are spectacular cases.

It is easy to see why governments love Facebook, Twitter and YouTube. The network effort ensures wider reach of their content, precisely because ‘people are on Facebook. Take it where they are’. It is also relatively cheaper resource-wise to use platforms because the overheads may just be data costs, a social media manager, and remembering the password(s). What’s not to love.

However, the resources supposedly saved from such a strategy are paid for dearly in its implications on content management. For starts, platforms are private entities whose rules or terms of use may be in contradiction with government strategies. Some companies have dedicated officials to handle government requests but it sounds ineffective to have to go through half dozen hops to communicate with your citizens. An example would be a government posting content on platforms that require membership. Streaming an official government event on FacebookLive may sound dirt easy, until you realise that your citizens have to signup for Facebook accounts else their experience is highly diminished compared to those logged in. Further, future content reliability depends on platform regulations and availability, most of whom operate closed and proprietary software, limiting transparent audit of what happens to the content. This results in government communication, a core pillar of democracy, being overly dependent on the rules of software developers and policy wonks in far off lands. But this need not be the case.

Publish on Own Systems, Syndicate Everywhere

There is a tried and successful way out to leverage platform reach without sacrificing government communication strategies. Publish on your sites, syndicate everywhere (POSSE). In case of contradictions, your domain is the correct version. Using free and open source platforms self-hosted in their jurisdictions, African states can take back their control of information cycles. A working example of such a system would be a citizen-focused website paired with a self-hosted Mastodon instance run and maintained by the government. All official communication is published on such a platform, linked to all departments, and secured as any other State infrastructure. This calls for proper staffing, healthy information and archival management, and most importantly, policy makers who understand government communication as a pillar of good governance. These are priorities a serious government would be willing to invest in to maintain a sense of communication sovereignty.

It may be reasonable to argue against this by claiming that even if government departments publish on their accessible platforms, the ‘average’ Internet user will not necessarily dig through all the text to verify information. However, it should not be lost on us that journalists, fact-checkers, researchers, and public interest technologists will have an easier time reporting or verifying, which in turn improves the general information hygiene in society.

Platforms may be the reality we live with for now, but they need not be our everything. As concerns around their disproportionate role in mass manipulation gets increased attention, we should not forget the main mover of public information. Without verifiable information, anything goes. This is the perfect ground for disinformation. It is time to bring the State back into the conversation.