The art and practice of intelligence design

Or what are collective intelligence assemblies and why are they the game we must learn to play so we can play the game?

Big Mind: How Collective Intelligence Can Change Our World https://press.princeton.edu/titles/11098.html Geoff Mulgan, Princeton University Press 2017

“Better together” increasingly goes the mantra when the conversation turns to solving big problems whose characteristics combine complexity and scale with intensity and intransigence.  Shifting models and practice of school education from the ground up, finding better and more affordable ways to connect health and social care, making our cities more liveable or working out how to adapt to, perhaps even to avoid, catastrophic climate change. You know the sort of thing.

It’s not hard to add to the list…shifting thick patterns of working to accommodate the rise of the robots and the ambiguous march of artificial intelligence, lifting the trust quotient of large public, corporate and civic institutions that are increasingly out of shape and out of place in a connected world of speed, intensity and subversive legibility.

I’m not sure when it was ever a good idea to try to solve these kinds of problems by relying on simple or singular responses.  It certainly isn’t now.  The answers, if there are any, are going to come from our capacity to think, imagine and act together in more and more interesting and often unusual combinations of skill, experience, knowledge, decision-making and delivery.

In other words, our capacity for collective intelligence.

It’s the same problem again – we’re not short of intelligence, we’re just short of good ways to line that intelligence up to larger collective outcomes.  So even if it’s true that we’re “better together”, it turns out that getting together is harder than it looks.

Big Mind, Geoff Mulgan’s new book, is the culmination of 20 years of thinking, advising, designing and delivering on the front line of activism, policy and public work in the UK and around the world.

It’s part manual and work book, part philosophical tract, part invitation to learn and part call to arms for the growing practice of our ability to think and act together, “the capacity of groups to make good decisions—to choose what to do, and who to do it with—through a combination of human and machine capabilities.”

He argues from the early part of the book that “the ways intelligence is organized are largely fractal in nature with similar patterns occurring on multiple scales, from groups of friends to organizations and whole societies.”  If you remember from your complexity theory, “fractal” means that the shape and patterns of organising are the same at whatever scale you observe the phenomenon you are studying. Think sea shells, snowflakes, lightning, broccoli, peacocks and pineapples.

“On each scale,” Geoff argues, “collective intelligence depends on functional capabilities: distinct abilities to observe, analyze, remember, create, empathize, and judge—each of which can be enhanced by technologies, and each of which also has a cost.”

These are then supported by infrastructures that make collective intelligence easier: “common standards and rules, physical objects that embody intelligence, institutions that can concentrate the resources needed for the hard work of thought, and looser networks and societies of mind.”

There’s an important clue in that description to one of the consistent themes of Geoff’s work more broadly. The ‘dynamism” of collective intelligence relies on a “mechanism” of large institutions together with looser, and usually digitally mediated networks of knowledge and practice, to actually do the work.

And that work is always the same – concentrated, hard and often unspectacular thinking and analysis, communication and diffusion, alignment through incentive structures (which are often either unhelpful to the common cause or positively antagonistic to it) and delivery.

Increasingly, he suggests, these collective intelligence infrastructure or capabilities “are hybrid combinations of machine and human intelligence.” These assets are built on the principles of continuous learning and reflect five characteristics:

“…they create autonomous knowledge and informational commons, achieve an appropriate balance between their functional capabilities, achieve focus, orchestrate systematic reflection, and integrate for action.” (Emphasis added)

But there’s a problem, obviously.

“Powerful tendencies in organizations and societies—including conflicting interests—push in opposite directions to each of these, which is why they are rare.”  In other words, the way our institutions and systems are currently configured to push and shove in pursuit of their ambitions is often antithetical to the interests of a collectively intelligent outcome.  This is the phenomenon, which Geoff calls out early in the book, that often confronts us with individually smart pieces of a “dumb” system (or how it is possible for organisations sometimes to act in ways which are egregiously stupid despite overflowing with individual smartness, even brilliance.)

The book then takes an important tour of the world of “triple loop” learning:

“The everyday processes of intelligence then operate at multiple levels that link together in a hierarchy: a first loop using existing models to process data, a second loop of learning that generates new categories and relationships, and a third loop that creates new ways of thinking. These can be combined in triggered hierarchies.” (Emphasis added)

Collective intelligence works best when it activates all three loops and levels, each of which is triggered when the opportunity or risk we’re trying to tackle demands that other, larger and higher parts of the system become engaged.

This is how the book describes the alignment:

“Groups and organizations think well when they have all these in place, with a balance between capabilities, effective infrastructures, systematic ways of managing the three loops, and a willingness to devote resources to the hard work of structured thought, to tap into a bigger mind beyond their own boundaries and remain self-conscious about methods. But most important fields of human activity lack crucial elements—and so end up much more collectively stupid than they could be.”

Hands up anyone whose organisation has a tendency to “end up much more collectively stupid than they could be”? I thought so.

Geoff’s first academic and research interests were focused on communication and networks, so it’s not surprising he makes (appropriately) much of the role of the Internet itself.

“The spread of the Internet along with ubiquitous tools for analysis, search, and memory”, he argues, “have greatly enhanced the world’s capacity to think.” The problem remains that “many more resources are devoted to collective intelligence in competitive fields than cooperative ones, however, and the world suffers from a huge misallocation of brain-power….”

It’s the same problem again – we’re not short of intelligence, we’re just short of good ways to line that intelligence up to larger collective outcomes.  So even if it’s true that we’re “better together”, it turns out that getting together is harder than it looks.

Collective intelligence assemblies

The book, in line with Geoff’s more recent writing for Nesta on his collective intelligence ideas (here and here) gets more interesting when he describes how “collective intelligence assemblies” are created.

He suggests that the successful examples of collective intelligence “are best understood as assemblies of multiple elements. Discovering which assemblies work best requires continuous shuffling of the elements, since capabilities, infrastructures, and organizational models have to coevolve with environments.”

That makes intuitive sense.  What we’re trying to do is render the pieces – knowledge, expertise, physical and virtual assets, platforms, money, policy, power and authority – more intelligent as they assemble themselves in something that starts to look like a purposeful and decisive puzzle.

Yet, he continues, “some of the most important fields—including politics, the university, and finance—lack this capacity for iterative shuffling, and so become locked into configurations that keep them less effective than they should be.”  This is crucial.

As I argued in a small book a few years ago written with former Finance Minister Lindsay Tanner, institutions too often find themselves out of shape for the new roles of collective intelligence (as Geoff offers the framework) to which we want them to contribute.  As Lindsay and I argued, “changing shape” is an institutional reform imperative that now looms larger than ever.

Geoff’s focus is as global as it is ferociously and predictably local.  At the global level, he suggests, “there is a need for new assemblies that can marshal global collective intelligence for global tasks, from addressing climate change to avoiding pandemics, solving problems of unemployment to the challenges of aging.”  He think it’s possible to imagine what these could look like—building on recent initiatives in medicine and the environment that try to observe, model, predict, and act.

These are three examples taken direct from the book:

Google Maps

The spread of Google Maps is a good example. It started off with a grand ambition of organizing global geographic knowledge in a comprehensive and usable form. But Google lacked many of the crucial skills to achieve its ambition and so brought in—or to be more precise, bought in—others: Where 2 Technologies, a company founded by two Danish brothers, which provided a searchable, scrollable, and zoomable map; Keyhole, which developed the geospatial visualization software that would become Google Earth; and ZipDash, which provided real-time traffic analysis, based on information gathered anonymously from cell phone users.

This assembly of different elements supplied the spine for a truly global system of geographic knowledge. Next Google had to tap into a much wider set of skills to make the maps more useful. It did that by opening up the software—through the Google Maps API—to make it as easy as possible for other sites to integrate it.

Platforms

Some of the most interesting hybrid assemblies use platforms to aggregate and orchestrate ingenuity on ever-larger scales. Makerbot Thingiverse is a platform that hosts over 650,000 designs for the maker movement. WikiHouse shares elements of design for anyone to design their own house—encouraging users to put their own adaptations and ideas back into the commons.

In health, a new family of platforms allows people suffering from acute conditions to become collective researchers, turning a disparate group of patients into something more like a collective intelligence by using a mix of digital and human thought. Current examples have recruited people with Parkinson’s disease and supplied them with wearable devices with accelerometers so that they can pool data about how they were faring, and a similar approach has been applied to dementia.

Cancer

One of the most comprehensive existing collective intelligence assemblies quietly supports cancer treatment in England’s National Health Service through the National Cancer Registration and Analysis Service.

Despite the mundane name, this is an extraordinary feat of organization that points to how many public services and whole systems could be run in the future.

It links thousands of records—including the three hundred thousand new cases of cancer in England each year. It brings together diagnoses, scans, images, and past treatments. The data feed into predictive tools to help patients choose different treatment options.

Where necessary, the data are linked to genetic information, or other data sets that help to predict if the illness could lead to debt or depression, and market research information to help better target public health messages. The whole array of information is then used to guide the day-to-day decisions of doctors and increasingly patients too.

In the end, the book is an extended tussle with something I suspect Geoff has been wrangling most of his professional life as an activist, government worker, advisor, strategist, innovation advisor and investor and practitioner – “how do societies, governments, or governing systems solve complex problems, or to put it another way, how do collective problems find collective solutions?”

And many of us will know why that deceptively simple invitation is, well, deceptively simple.

For example, “within any group, diverging and conflicting interests make any kind of collective intelligence both a tool for cooperation and a site for competition, deception, and manipulation.”  Collaboration, Geoff reminds us, can cut both ways and is not necessarily a “good” in its own right.  It is a way of getting work done.  The question is, what is the work?

Similarly, collective intelligence can find itself strung out between “the silence of the old hierarchies in which no one dared to challenge or warn, and the noisy cacophony of a world of networks flooded by an infinity of voices.”  Too much silence or too much noise will kill the instinct to assemble and use multiple intelligences.  As always, the (necessarily experimental) search is for the sweet spot, recognising that the spot keeps moving and only reveals itself as a function of persistent search, learning and action.

Collective intelligence, as Geoff describes it later in the book, “can be light, emergent, and serendipitous.” More often, though, “it has to be consciously orchestrated, supported by specialist institutions and roles, and helped by common standards. In many fields no one sees it as their role to make this happen, as a result of which the world acts far less intelligently than it could.”

Quick (self-interested) side bar…

It is precisely to experiment with new ways to fill that gap – the gap left by the absence of groups or process that “see it as their role” to do this kind of purposeful connecting and intelligence-directing – that The Impact Assembly, a new venture from PwC’s social impact practice has been established.  I am working with the Assembly (at least we picked the right word!) to find effective ways to ‘assemble’ disparate pieces of collective action puzzles into useful and disciplined collaboration.

Current pieces of work are happening in domains as diverse as obesity, mental health and suicide prevention, young people and the new work order, multiple sclerosis and homelessness.

Geoff is right.  The stakes could not be higher.

Progressing collective intelligence “is in many ways humanity’s grandest challenge since there’s little prospect of solving the other grand challenges of climate, health, prosperity, or war without progress in how we think and act together.”

It is, in that sense, the ultimate enabler and, for that reason, a rising target for innovation and invention in its own right.

It reminds of earlier definitions of social innovation than come from research and thinking from Geoff and others, especially from The Young Foundation, captured in this definition from a major review of the theory and literature:

 Social innovations are new solutions (products, services, models, markets, processes etc.) that simultaneously meet a social need (more effectively than existing solutions) and lead to new or improved capabilities and relationships and better use of assets and resources. In other words, social innovations are both good for society and enhance society’s capacity to act.

Like social innovation in this definition, you get the strong sense from the book that collective intelligence has the same dual meaning – getting problems solved or at least tacked productively and, as a consequence, building the capacity to solve future problems.

Like all good habits, I guess, collective intelligence gets more useful (for the problems you want to solve in the future) the more you use it (to solve the problem you’re trying to solve now).

In a way, I suppose, the idea of collective intelligence could be just a fancy way of reminding ourselves that, in whatever group we find ourselves – a family, an organisation, a city, a country or the globe – we’re almost always more than the sum of our parts.  There is often a wisdom in the crowd, a way and strength of knowing and acting that gains immeasurably from its social or networked provenance.

And in a world grown too big to know, as David Weinberger so elegantly put it, the smartest person in the room is very often the room. Geoff’s collective intelligence thesis is all about how we design tools, cultures and assets that turn what the room knows into purposeful and productive action.

In that sense, collective intelligence points to the shifting nature of the kind of knowledge we need to survive and thrive in a world in which speed, intensity and a kind of radical connectedness renders the business of turning knowing into useful acting irreducibly collective.

“Assemblies are in part technical designs, but they only become useful if they connect to action, which requires them to be sophisticated about behaviors, cultures, and organizational norms, all of which may be more taxing than the design of sensing systems and algorithms.” Geoff Mulgan

Geoff’s work oscillates, often at great speed, between theory, action, learning, adjusted theory and more action.  In the book, he complains that “these ways of organizing thought on a large scale are still in their infancy.”

As a consequence, he admits he’s writing about ideas and potential practice that “lack a convincing guiding theory and professional experts who know the tricks of the trade. In many cases, they lack a reliable economic base. Yet they suggest how in the future, almost every field of human activity could become better at harnessing information and learning fast.”

This is the book’s call to arms, which is in effect a call to learn, together.  It constitutes a broad and inclusive invitation to join the effort to start filling out the collective intelligence playbook.  And that playbook should reflect that collective intelligence theory has to be constructed, at least in large measure, “not just by interests and habit but also by meanings and stories.”

And it has to engage directly this uncomfortable paradox, that “the very properties that help a group cohere can also impede intelligence.”  What are they?  Geoff lists these: “…shared assumptions that don’t hold true, a shared willingness to ignore uncomfortable facts, groupthink, group feel, and mutual affirmation rather than criticism.” Often, shared or collective thinking “includes not only knowledge but also delusions, illusions, fantasies, the hunger for confirmation of what we already believe, and the distorting pull of power that bends facts and frames to serve itself.”

It turns out that “most groups face a trade-off between how collective they are and how intelligently they can behave.”

As Geoff summarises it, “the more they bond, the less they see the world as it really is. Yet the most successful organizations and teams learn how to combine the two—with sufficient suspension of ego and sufficient trust to combine rigorous honesty with mutual commitment.”

High trust, low ego?  That shouldn’t be difficult…

This is a book that needs a bit of time to study and ponder. I keep going back to chunks of the argument and share the narrative with colleagues and clients alike.  As always with a Mulgan book, there is plenty to challenge and provoke.

The book is not the answer, or even an answer, but it does have answers that test and stretch.

It draws on more than 20 years of intense study and practice to offer some sharp, difficult but very valuable questions. The American writer Thomas Berger one said, apparently, that the art and science of asking questions is the source of all knowledge (he also, even more pithily, answered the question “why do writers write” with the obvious answer, “because it isn’t there.”; not strictly relevant here but a profound insight worth sharing).

If knowledge and questions go together, Big Mind is full of knowledge we can usefully engage, test, use and perhaps even add to.

In the end, the book is dedicated to the proposition that “creating such [collective intelligence] tools on a scale, and with capabilities proportionate to the challenges, and nurturing people with skills in “intelligence design” will be one of the great tasks facing the twenty-first century.”

I think’s he right.