A Manifesto for Public Technology

If democratically elected governments are going to be able to meet their citizen’s desires to limit and control the activities of digital companies, then a new class of public servants are going to have to be trained up to do that work.

Let’s start with a provocation:

The world’s national, state and local governments don’t have the right digital skills in the right quantities to meet the challenges of the coming century. This is a Big Problem.

I’m calling this a provocation rather than a statement of fact because, despite looking, I can’t really find anything approaching reliable data on the quantity or quality of digital skills spread across governments. Read into that data gap what you will.

If you don’t share my suspicion that governments are understaffed to cope with the digital age, then you can stop reading now.  I’m sure there’s some other explanation as to why Healthcare.gov blew up and brought Obama’s approval ratings to an all time low, and why Hillary Clinton’s chance of being President of the Goddamn USA is most seriously threatened because of a Goddamn email server. And I’m sure that with a wide public-sector track record of cocking up relatively simple digital projects that there’s virtually no chance of something truly catastrophic happening the first time governments try to seriously regulate genuinely difficult digital things like Facebook’s wall algorithm, or self-driving cars’ ethical logic.

However, if you do share my theory that there’s a seriously problematic skills gap when it comes to governments and digital technology, do please read on. The rest of this post is about things we can do together to make sure our governments get the skills they need.

We need a public technology hippocratic oath

Ubiquitous sensors and oceans of data produce a range of ethical dilemmas that will trouble politicians and public servants for the foreseeable future.

It simply isn’t reasonable to ask harried public service workers, or even harried politicians, to discuss, from first principles, the fathomless moral conundrums related to data and technology every time a new digital service is set up or modified. People need moral frameworks to lean against—always have, always will.

Consequently, a new set of techno-ethical principles need codifying that are easy to remember, easy to teach, easy to hold people accountable against, just as the principles are that are taught to doctors or soldiers alike. They need to be the sorts of things that are drilled into the heads of students, sworn on during appointment ceremonies, and stuck on the back of office toilet doors for reinforcement later in life.

This authoring process is going to mean an intellectual battle royale of digital thinkers and digital principles. I consider preparing and defending such sets of principles a practical challenge for the likes of Zeynep Tufecki, Evgeny Morozov, Tim O’Reilly, Mustafa Suleyman, the new center for Data & Society, and pretty much all of the world’s academic moral philosophers. And when it’s all over, it’s highly likely that no single idea or set of principles will have won. But at least there’ll be a range to choose from, and some frameworks for people to fall back on, rather than none.

We need universities to establish new centers for research on and teaching of public technology

In the MOOC era it’s almost embarrassing to be a professional technologist talking about the need for academic teaching that takes years and is situated largely in one building. But I’m going to take one on the chin for bricks, mortar, and non-pixelated academics.

The reason we need to talk about new academic centers and new academic teams as serious, physical ventures, is that the amount of work that needs doing in the new public technology sector is just enormous. We need to produce and stress-test lots of new ideas, we need to produce lots of new people, we need to produce lots of new data and research.

There’s just much too much work to be done than can be managed by slotting into the cracks within an existing research center or an existing master’s program. There was no way that all of nuclear physics research could be squeezed inside the spare rooms of old 19th century physics faculty buildings, and the same will be true of public technology. The issue is too big, and the amount of outputs required are just too great.

[UPDATE:] The good news is that there are already promising signs of movement. The GovLab at NYU doesn’t just publish papers but has been actively coaching public servants to push through a huge number of new digital government projects—they claim an impressive 250 in the last year. Elsewhere in the same university CUSP is focusing on data at the city level, especially how it is acquired and exploited to help cities become more ‘productive, livable, equitable and resilient’. It has codified knowledge into an actual masters program—the MsC in Applied Urban Science and Informatics that points the way towards how more formally recognized graduates can emerge from this field. And the action is not just reserved to New York—in Princeton the Center for Information Technology Policy works less on city-level delivery projects and more on the ever-growing world of pure policy questions that the internet seems to generate every day. Berkman and the MIT Media Lab have touched on this field for years. And the Center for Government Excellence at Johns Hopkins is going great guns too.

What I hope is that the people running universities (and the people funding universities) become ever more sympathetic towards the claim that there is a significant public interest in working on public technology ideas and skills. A start has been made, but the number of actual certified courses that wannabe public technologists or general public servants could join today is relatively few, and the range of formally recognized qualifications relatively modest.

The rest of this piece essentially sketches out various programs of work that could be tackled by this sort of new centre for public technology, and for similar but non-degree-granting institutions such as think tanks, companies and dedicated public-interest NGOs like Code for America. Geoff Mulgan recently mooted the idea of a machine intelligence commission; this idea represents exactly the sort of non-university institution that will also have a part to play in fleshing out the whole public technology space.

We need a lot of new thinking about what good public technology looks like

Before we can get really good at teaching a lot of people to be public technologists, we need to develop a lot of new ideas about what good public technology actually looks like. The Government Digital Service and 18F have already started strongly on this front, but their relatively modest size, age, and remit means they’ve only just started to scratch the surface of all the thinking that needs doing.

Ultimately good thinking needs to emerge on issues varying from ethical ideas (see above), to procedural ideas (how do we get more user-centered design into governments?), to very difficult and very technical ideas (how does a government audit private-sector machine learning algorithms for unintentional-but-emergent sexism?).

This is going to need quite a lot of people given quite a lot of space and time to look into very different issues. Public sector procurement reform couldn’t possibly be more different to auditing the ethics coded into self-driving vehicles, but both will need serious brainpower if our governments are going to serve their citizens effectively. This is going to need a lot of new people, a lot of coffee, a lot of conferences, a lot of papers, a lot of real world experiments, and a lot of publications. Public Technology as an issue is so big that it represents a whole new school, connected with and dependent upon many other schools, but separate and sizeable in the same way that medicine and the law are separate from general schools of government.

The ideas are, of course, only important if they make it into government, which is what the next section focuses on.

We need to train up a lot of specifically public technologists

All of the ideas generated through the above intellectual ferment will remain firmly excluded from government itself if we don’t create a whole new generation of public technologists who can infiltrate the governments of the world and bring change from the inside. These individuals will have to have extremely strong, Silicon Valley-grade technical skills, but also a better-than-basic grounding in the history and operation of the state, the legal system, and the quirks of democratic structures. Together they can get stuck into the problem that, in rich countries, cities alone deliver hundreds of services each, almost none of which have been transformed in the way that so many private sector services have been. There is a giant backlog of work simply to make the public sector even resemble the private one, in terms of user experience and efficiency.

We need these people for many reasons, though, only of which only the most basic is “making sure government websites work OK.” A need every bit as vital as running the government’s own systems is going to be the problem of regulating digital systems run by other people.

If democratically elected governments are going to be able to meet their citizen’s desires to limit and control the activities of digital companies, then a new class of public servants are going to have to be trained up to do that work. And because even one algorithm inside one internet giant is so unfathomably complicated and so unfathomably wrapped up in legal protections, we’re going to need really quite a lot of these people.

There is currently an entire missing generation of BAs, MAs, and PhDs who are not entering the public service at all, let alone with the right skills. But widespread public demands for both better government services and more control over digital companies are going to make these skills shortfalls less and less tolerable. Ultimately it’s going to be the universities that will be the key actors in reacting, followed by professional training organizations. And a lot of care will have to be taken to make sure that this training doesn’t replicate in government the gender and ethnic inequalities that plague the technology sector.

We need to influence the university and professional training of mainstream politicians and public servants

If the decision makers of 2050 have a new cadre of public technologists below them, but have zero understanding of digital issues themselves, the problems of digitally ignorant governments will not go away.

Digital technology issues are different from some of the more familiar techno-political challenges of the 20th century. As a political leader you don’t have to understand nuclear physics to understand what mutually assured destruction means for your foreign policy. As a health minister you don’t have to know much about the interactions of drugs with cells to know that a pill of cost X that stops Y people from needing costly medical treatment Z is worth it.

But digital technologies represent a change to the fundamentals of government, society and economy that’s more analogous to the rise of economics than the rise medicine. A lack of digital skills will ruin a governor’s ability to do their job, in the way that a lack of medical knowledge won’t. After all, if a big chunk of your money supply is running on some sort of blockchain-based system, can you make good fiscal policy if you have not even the faintest idea what that means? And if your civil servants tell you that it’s time to replace some manual public services with similar services partly powered by a new form of AI, how can you be a good leader if you don’t know what sorts of questions you should ask? And what if people start passing laws that are themselves in some sense algorithmic—will you be able to vet the bill?

So, as well as creating a new generation of public technologists with deep technical skills, we also need to create a new generation of mainstream public servants and politicians who have a similar amount of technology knowledge as they have in more common areas like political philosophy, history, or economics.

To be clear, these will not be political and public service leaders who can code entire applications or hack into the computers of foreign governments (although that would make for a fun diplomatic crisis). They will instead be public service leaders who are equipped with a mental toolbox of computing concepts and key questions to ask, as well as strong familiarity with the ethical principles discussed above. They won’t be computing experts, but they will be literate of key computing concepts.

This development will require widespread collaboration between the new centers of public technology expertise and the pre-existing university departments of political science, social science, literature, history and business that supply so many of the world’s public servants and politicians. There will definitely be tricky moments, especially the part where the news is broken that some older skills probably have to be dropped to make room for the new: I was taught metaphysics, for example, in my own degree-for-political-leaders.

We need more data on how many public technologists are needed, and for what purposes

We need a lot more data about the scale of need for public technology skills, the scale of production of those skills, and their impacts once they are delivered.

As I stated at the top of this piece, I can’t myself provide quantified evidence on the need for public technologists, in the way that I could for well-documented teacher or doctor shortfalls in many countries. Those of you still reading probably agree that a shortfall exists, but you probably can’t quantify it either.

That degree of shared suspicion is more than enough justification to spend some initial money on getting started, but it isn’t enough to see training and research reach a really large scale. Evidence will need gathering on both need and effectiveness of public technology teaching before public technology becomes a skill set that is more than niche and experimental.

New university centers are ideally positioned to work with concerned governments to map the scale of the problem, and thus map the scale of response needed. Does the U.S. need two Master’s programs producing pubic technologists, or two hundred? This is a question that can be answered, but only through diligent work and close collaboration.

What next?

This piece is not supposed to be a bombshell dropped on an unsuspecting world. I am writing it because it is a consolidation of things I have heard a great many people say, as well as my own ideas and, of course, prejudices. But writing it down under one heading I hope to provoke people who see something in this to come forward. Citizens of the 21st century need public technologists like citizens of the 19th century needed municipal engineers to build the drains and clean water supplies.

As societies, we’ve done this once before. It’s time to talk about whether we can live up to our own ancestors.