Designing the Office of the Social Media Regulator — a blueprint
I estimate a regulator would cost £300m per year to fund and need 3,300 staff.
Social Media Regulation is a current topic with politicians:
- Theresa May says the internet must now be regulated following London Bridge terror attack,
- Angela Merkel calls for transparency saying internet search engines are ‘distorting perception’, while
- Xi Jinping has been regulating for a while, and is getting stricter: Xi Jinping ramps up his crackdown on the Chinese media — both online and off
The arguments for social media regulation are mounting:
- Social Media is a public utility that is essential for modern life and so must be regulated to ensure minimum service for all citizens.
- We need accountability in how algorithms work — particularly how we are recommended content — does Youtube’s recommendation algorithm channel you to increasingly harder core content? does Facebook’s newsfeed keep you in a filter bubble? how does Google really decide who comes top of the search results?
- Reduction of fake news — does the business model of many social networks wrongly incentivise the production and monetisation of fake news?
- Improving behaviour — such as reducing cyberbullying and trolling
- Maintaining privacy — controlling who can use our personal data and when?
- Cutting out spam — bots that cheat the algorithms and drown out the real conversation, so wrongly highlighting the most important news and people
- Policing communications —providing law enforcement and security services appropriate access to conversations and activity by criminals and people of interest
- Judicial process — enforcing existing laws in new mediums such as taming the online lynch mob with digital justice.
- Fair advertising — maintaining advertising standards, particularly around election time with micro-targeted dark ads offering here-today-gone-tomorrow promises.
- Digital rights — providing fair process in disputes — for example at the moment you can be barred from social media for a breach of terms with no right to appeal — the platforms act as prosecutor, judge and jury. This seems unfair.
That’s already a long list and undoubtedly there are many more areas to address.
Clearly there is a need for social media regulation, the question is whether it is more of the same (advertising, GDPR, libel and existing media rules already provide considerable protection) or whether a step change is needed to bring the strands together under one roof — I propose the formation of the Office of Social Media Regulator (let’s abbreviate this to OFSOC for now).
Before we dive in further, let’s make sure we recognise the arguments against further regulation — after all, the only thing worse than no regulation is bad regulation:
- Regulation could stunt innovation for existing firms so giving new entrants an unfair edge.
- Regulation could make the current de facto monopolies de jure and actually prevent future competition.
- Innovation occurs too rapidly to regulate. By the time you’ve regulated a firm or a feature it has already been superceded by someone or something else.
- Regulation could empower authoritarian regimes to enforce easier state censorship — so harming the spread of liberal values and potentially condemning thousands of bloggers and journalists to prison.
- It is really the duty of providers to regulate the users of their services by hiring more moderators for example, not something that should be a burden on the tax payer.
Again a growing list and I am sure there are more arguments to be made here too. The point is that the lines we take on social media regulation are not going to be clear cut and easy to solve. There is plenty of thorny debate still required.
But let’s imagine that the case is made, agreed and we all want to progress with Social Media Regulation. What now? How do we do it?
How would we design OFSOC?
I think there are a few aspects to consider:
Size
Social Media is a monster — 2bn people on Facebook are not going to be effectively regulated by a small group of well meaning bureaucrats in a third floor office in Brussels.
You don’t set a mouse to catch a cat. OFSOC will need global muscle and reach.
Speed
The internet works fast, to keep pace, a regulator must be fast too.
This will increase the cost of OFSOC.
It’s worth bearing in mind the traditional project management triangle — you can have fast and high quality but you can’t have low cost. Or you can have fast and low cost but you won’t have high quality.
Unfortunately this is public sector so quality is a necessity. Making lots of mistakes in executing regulation won’t help anyone.
Global
Since digital and social media is inherently global there is no point creating a regulator that only regulates at the national level. It simply doesn’t make good sense.
It would be like trying to regulate a river by only looking at the section of the river on your land — it’s just not possible. You have to regulate the whole cyber river, no matter whose land it flows over.
This means that the governance of the organisation would need to be confederalised (a fancy word meaning voluntary handing over of power from the edge to the center) in the same way as nation states sign up for the World Trade Organisation or the United Nations.
Scrutiny
Any organisation wielding power needs to report in somewhere and be scrutinised. A scrutiny committee would be needed to monitor the actions of the regulator, in turn reporting to publicly elected officials. The question again would be where the people doing the checks and balances sit.
Now assuming we’ve dealt with issues of funding, establishment and governance, how do we actually run an OFSOC?
I think the best way to do it would be to organise functionally first, with national outlets for each function only as needed.
There will be a need for a global headquarters — let’s put that in London for now. I appreciate it could go anywhere.
The traditional UN organisation approach of having a self standing local missions tends towards the creation of national differences. Our objective is to regulate social media as a whole, not to create separate national digital fiefdoms.
Instead I would prefer to organise functionally and then each function create local missions only as necessary to fulfil their function in each country.
So with that mind, what are the functions of OFSOC?
Platform relations
Any good regulator starts by maintaining relationships with those it is seeking to regulate. This function would likely be based in Silicon Valley, on the doorstep of the head offices of the major tech firms.
This would ensure that the true decision makers at platforms (often engineers), rather than their local national counterparts (often PR people) are the ones we have a relationship with.
Any fines applied (a watchdog must have teeth) would be managed by this function.
Feature Review
This function would require investment to provide oversight of new features and assess regulatory impact.
With the incredible number of new features launched every day this would require someone following product hunt and every platform blog and each new feature reviewed.
What you say? review every feature — nonsense and impossible! Well maybe but every media feature can have an effect — some argue that the Facebook Safety feature actually plays into terrorist hands by personalising and publicising an incident more widely than needed (Why I won’t be marking myself safe on Facebook today).
This function would need to offer both open and closed hearings — for example some features will need to be kept private to avoid gaming the system (new tweaks to the Google search algorithm for example) that should be accountable but not transparent.
Judicial Process
There is a need to coordinate day to day actions and treaty agreements between the justice system, law enforcement and the platforms. This function would ensure a smooth flow of information between all the parties.
Public Communications
Changing behaviour online requires a commitment to public communications. There is no point in underfunding this as changing demand is the best way of reducing unwanted supply.
With a constrained public purse, public communications budgets have gone a bit out of fashion, however any good organisation knows, if you want your message heard and acted on, you need to spend money on advertising.
Improving social media etiquette will go a long way to countering issues of lynch mobs, trolling and cyberbullying.
Public Policy Partnerships
The regulator must also handle the interface between social media regulation and national public policy. Inevitably there will be cross over and a need to join the dots.
Citizen Rights
A good regulator should defend citizen rights and social media is no exception. Providing third party arbitration an appeal process for the disputes between social media users and their platforms will be vital.
Report Arbitration
When the platform fails to take down content after a reasonable time, who has the power to intervene and force a take down (or indeed force a restoral) ? It will be the regulator. With 6 million reports per day on Facebook alone, many of these will need a dispute resolution process as no moderation process is perfect.
New Platforms
Closely associated with the feature review there may be a need to regulate the creation of new platforms — do we cut off at source apps that are clearly harmful? Examples of this include people rating apps — Black Mirror’s horrific people-rating app is now a reality. Sort of, Peeple App Will Let You ‘Rate’ Your Friends With A Star Rating
Secretariat and Senior Management
There will of course need to be an internal function to handle all the staff, infrastructure to support the various functions, as well as a senior management team to oversee smooth running.
Department for everything else
This list of functions is a start — there are undoubtedly more I’ve not thought of!
How much will it all cost?
If we put a head count in each function, multiply by the number of regulated platforms, citizens and countries then we can perhaps estimate a cost for staffing and running OFSOC.
I’ve done this exercise myself and I estimate the organisation size and budget as follows:
Head count
Head office: 1532
Regional offices (100): 1800
Budget
Salaries: £266m
Expenditure: £34m
So in total I estimate OFSOC will cost about £300m and require a staff count of about 3,300 people. What an interesting exercise.
I wonder if it will ever be a reality?
