A thorough and more rounded understanding of collective intelligence matters for obvious reasons: to help avoid unnecessary errors and disasters and to benefit from the explosion of new digital technologies. What does the future hold for both collective intelligence and collective stupidity?
The world is packed with examples of collective stupidity. Just look at the US political system – a nation full of brilliant people yet decisions are made in ways that rarely reflect its capacity. Financial markets are another prime example – full of brilliant people unable to learn the lessons of history and narrow-minded when it comes to judgement and ethics.
Interesting empirical research has examined what factors influence the collective intelligence of a group or institution. It pointed to three chief characteristics: the average social perceptiveness of the group members, relatively equal turn-taking in conversation and the percentage of women in the group. Beyond this, however, there has been little rigorous probing of what makes a group or institution smarter.
To arrive at better answers, we need to distinguish the crucial components of intelligence. A few of the key elements are as follows:
: the ability to see, hear, smell the world
: the ability to focus
: the ability, or abilities, to think, calculate and reason
: the ability to imagine and design
: the ability to remember
: the ability to observe one’s own thought processes
: the ability to make another an end not a means
: the ability to judge
: the ability to make sense of complexity and to integrate moral perspectives
We all know technology is advancing rapidly, but much of what we are learning about human intelligence emphasises the importance of intuition, short-cuts and emotions in allowing efficient judgements to be made.
Immediately, it is obvious that some of these functions can be performed very effectively by computers and related technologies. Machines have vastly enhanced our ability to observe, calculate and remember and their capacities have far outstripped our ability to keep up. Others functions are, so far, largely untouched by technology, such as capacities to create or judge. We all know technology is advancing rapidly, but much of what we are learning about human intelligence emphasises the importance of intuition, short-cuts and emotions in allowing efficient judgements to be made. Machine learning may help or mimic these, but so far there are few very good examples.
Observation of intelligent people and organisations reveal loops of reasoning, existing in a rough hierarchy of three orders (think of them like levels). First order reasoning involves the application of thinking methods to definable questions; understanding, reasoning with categories, problem solving and calculation. Second order reasoning involves the ability to reflect on goals and to create new categories where old ones no longer function. Third order reasoning involves the ability to reflect on capacities to reflect and adapt.
Being capable of only first order reasoning may appear dumb, even if in other respects it is very clever. As you might imagine, computing has proven better at first order tasks. This is why computers are powerful tools for playing chess but not for designing games. It’s why networked computers can help shoppers find the cheapest products and simulate a market but offer little help to the designers of economic and business strategy. Likewise, networked computers can help people onto the streets but not to run a revolution. This is territory where rapid advances may be possible.
Tools for collective use exist and are emerging – Twitter for news, Wikipedia for knowledge, Kickstarter for investment, eBay for commerce. Examples of more advanced forms of collective intelligence include platforms such as Couchsurfing and Buzzcar that connect users with underutilised resources, or reveal things that are practically invisible, such as Baby Come Home, (the Chinese site using facial recognition software to find lost children) and Blindsquare (the app helping blind people navigate cities). These platforms largely involve first order reasoning but other tools, such as multi-stakeholder dialogues, attempt second and third order reasoning, tending to require much more face-to-face interaction.
Alongside research, we need creative experimentation and evaluation to understand emerging tools and the best ways to use them. Where this field will head is unclear but interestingly chess may be a pointer. Years after the success of Big Blue in defeating Gary Kasparov, there is strong evidence that the best chess games ever played have been by human-machine pairs, where ‘freestyle’ chess players use computers to help. The future of truly intelligent, ethically informed judgement may look similar.
Geoff Mulgan CBE is Chief Executive of the National Endowment for Science Technology and the Arts (NESTA) in the UK and Visiting Professor at University College London, the London School of Economics and the University of Melbourne.