Government operating systems and government as collective intelligence.
Thu Mar 26 2015 00:00:00 GMT+0000 (Coordinated Universal Time)
Today’s continuing revolutions in digital technologies are constantly changing the options for how government can be organised, with new tools ranging from sensors and machine learning, to predictive algorithms and crowdsourcing platforms.
All governments depend on operating systems – a mix of techniques, processes and technologies that help them to do their tasks, from collecting taxes to catching criminals, educating children to building roads.
These can all be thought of as examples of the broader field of collective intelligence – concerned with how to mobilise multiple brains and technologies to help with the job of thinking.
To function well, governments need to observe, assess, plan, imagine, remember and judge, and during every era the structures and processes of government have been strongly shaped by the available technologies for doing these tasks of thinking.
Technologies can amplify the intelligence of every aspect of government – from democratic deliberation to financial planning, disaster management to public health
Roads to help communicate, rolls and tablets to record, councils to deliberate are all examples of this, and the great growth of government in the 19-20th centuries coincided with the spread of professions built around specialised knowledge; new tools like statistical surveys and technologies like the telegraph which all amplified governments’ capacities to think and act.
Today’s continuing revolutions in digital technologies are constantly changing the options for how government can be organised, with new tools ranging from sensors and machine learning, to predictive algorithms and crowdsourcing platforms. These technologies can amplify the intelligence of every aspect of government – from democratic deliberation to financial planning, disaster management to public health. They are contributing a denser informational aura around every activity – including traces, tracks and comments. Here I summarise a few of the possibilities for improving government collective intelligence over the next decade.
Possibilities for improving government collective intelligence
Greater awareness: accurate observation of the world has always been a challenge for governments. How to know the numbers of people in towns and villages; the scale of economic activity; or conspiracies against the state. The answers have included statistics (like GDP); surveys; intelligence agencies and surveillance, all offering facts about the world to help governments decide on threats and opportunities, and the issues that deserve attention. Today a stream of innovations are improving observation: new forms of data; sensors enabling new forms of regulation (for example of air quality, emissions from factories); automated aggregation of data such as citizen movement patterns taken from mobile phones; citizen generated data; and web-scraping tools (for example to make sense of emerging economic activities). The biggest challenge now for governments is how to cope with the volume of possible inputs. Computing power helps governments to analyse patterns. Key tools include predictive algorithms, which have long been used in healthcare (to predict whether a patient will come to hospital) and criminal justice (how likely that a criminal will reoffend). Larger scale predictions have been a lot less reliable, and economic forecasting remains as problematic as ever. But it’s possible that machine learning will help governments better plan for changing needs (for example, for healthcare). Other tools for analysis include ways of aggregating assessments, like Intellipedia which pulls together intelligence assessments from over a dozen US agencies.
The biggest challenge now for governments is how to cope with the volume of possible inputs.
Better ways to mobilise finance for impact: there was a burst of innovation around handling money 20-30 years ago, with the spread of new tools for raising money and managing it. Now there is a gap between the available tools and what’s needed. Some of the likely innovations of the next few years may include better means of applying investment methods to government actions on people (in health and education etc), linking preventive action to long term returns; better tools for making money transparent, disaggregating to understand cost structures; and mobilising finance for critical tasks such as innovation, including stage-gate funding methods.
More feedback: a series of innovations are being tested around the world to ensure better public engagement and feedback. Nesta’s D-CENT platform is being used in Finland to help cities make decisions, with APIs to enable the public to track and contribute at every stage; in Spain political parties are using the same platform to involve the public in shaping their programmes. In Paris 5% of the budget has been opened up for a participatory budgeting process. And everywhere social media are being used to support richer feedback. The methods for handling these inputs are far from settled – and some of these are a lot more labour intensive than the more traditional top down communications methods of the past.
Accessing wider sources of expertise: how to tap into a much wider range of sources of expertise - crowdsourcing platforms a rough first step towards more systematic use of experts, reputational devices. The Conversation – which now has over 20m hits a month – is a powerful way to tap the brainpower of academics to feed into current issues, and shows one route to answering this question. Platforms like Peer to Patent in the US which allows commentary on patent applications is another.
More empathy: much of the 20th century states’ growth involved roles in which empathy matters – doctors, nurses, teachers etc. This tends to be a blind spot for technologists and enthusiasts for new tools. But as Robert MacNamara – former head of Ford, the Pentagon and World Bank – pointed out, most of the mistakes states make derive from a lack of empathy. The idea of a relational state brings some of these ideas together – and points to governments paying more attention to the quality of relationships they have with citizens, and to the social networks that play crucial roles in everything from public health to the management of service delivery. Technologies can support this shift: tools like Social Network Analysis can reveal the reality of relationships – for example who helps who in a local policing system.
More creativity: governments need ideas, and better solutions. Here an array of new methods are being used to widen the range of sources, including: challenge and inducement prizes; crowdsourcing platforms; labs and teams; and accelerators. The craft of doing this well is still developing. But a common theme is the need to tap into citizen creativity.
Better memory: governments have always been built around archives, stores of past tax payments, treaties, laws, obligations. Digital technologies allow them to remember far more, with far easier access. ‘What works’ centres that bring together evidence on policy and practice are in part a way of organising memory more effectively. But overall the state of knowledge management within governments is low. Models like the Health Knowledge Commons point to a future where there is much more systematic orchestration of available knowledge – from very formal evidence to experience and citizen generated knowledge.
Smarter regulation: new waves of technology will require different models of regulation, for example of drones, personal data, access to genomic data and driverless cars. New principles of consent may be needed, along with new rules to determine who has access to data, and who has rights to inspect algorithms (though it will be hard, even impossible, to make sense of complex learning technologies). New technologies will also transform how regulation can be organised. Much financial regulation is already automated, just as banks decisions are made by neural networks, albeit with powers for humans to intervene when markets get out of control, and many areas of environmental regulation may be revolutionised by the intelligent sensors offered by the Internet of Things.
More collaborative collective intelligence: digital technologies in government have so far had most success in improving relatively simple transactions, automating tax payments, licenses, registrations, or enabling choice within services such as university applications. But technologies can also support larger scale collaborative problem solving, bringing together brain power across agencies and across sectors. Some existing models suggest the direction of travel. These include health collaboratives (used in the US and UK), orchestrating diagonal slices of particular fields, and combining attention to evidence, improvement and innovation within professions. The work on health knowledge commons suggests how these could evolve.
Local city collaboratives like the London Collaborative (which brought together the senior leadership, mobilised teams of officials and others to solve problems, and started to orchestrate academic and research expertise to contribute to public problem-solving) show how cities could think more effectively.5 At a micro level study circles within schools and hospitals; horizontal networks linking practitioners; and more formal learning systems like Project Oracle (for youth programmes in cities), suggest how intelligence can be organised more effectively.
Growing new cultures of government.
New technologies have co-evolved with changing cultures of government, just as happened in the past, as print, telegraph, television and computers changed how governments thought. These are a few of the emerging cultures that may thrive and may be the necessary complement to new hardware and software, valuing:
Visibility – of people, processes and results as the default.
Precision – with data, quantification and exactness in all things.
Quickness – moving to test and improve ideas in real world settings on a small scale rather than mainly paper processes for policy development and implementation of untested ideas on a large scale.
Interactivity – from TVs to buildings to institutions, leaving people less satisfied with any service or institution which offers no ways to interact.
Flows as the unit of action – digital technologies have evolved from the world of files, folders, desktops through the world of links and pages to the ubiquity of flows in the cloud, with intelligence as a cloud capability to be bought in.
Access rather than ownership, with the public sector developing its versions of KindleUnlimited, Netflix and AirBnB for such things as transport, libraries and education.
Authenticity, and walking the talk: officials trying out new methods on themselves.
Should there be more R&D for government itself?
Governments spend huge sums on R&D, often more than 1% of GDP. But they spend surprisingly little of this on their own needs. Half tends to go to the military and intelligence, and much of the rest is spent on projects defined by academics or business. But there have been few systematic R&D programmes directed to the types of function described above. Perhaps some more systematic R&D on the operating systems and technologies for government is long overdue, for example to test out what variants of machine learning are likely to be most useful and where. As preparation, a foresight exercise might look at maturing and emerging technologies and their most promising uses in the operations of government.
‘True Collective Intelligence: a sketch of a new field’, Geoff Mulgan, Philosophy and Technology, 2014 27:122-133
This report from a few years ago showed how SNA could map the patterns of partnership in a town or city, revealing the human reality of cooperation. Transformers, Nesta, 2008. These tools have yet to become widely used, but are relatively cheap and easy to implement. See my paper on public sector innovation, and the Nesta/Bloomberg Philanthropy paper on innovation teams and visit www.theiteams.org.