For a peer-to-peer approach to collective intelligence
It's eight a.m. on a Monday morning in 2007. In the Arcueil examination centre, a thousand heads crane with difficulty over wooden desks that are damaged by pens scratching across thin sheets of paper. Railway lines surround the enclave; trains make the building shudder rhythmically; the heads lookup for a minute, distracted, then return to concentrate on the studious, urgent writing of their paper. Invigilators wander the rows, imperturbable, watching every head that turns, every hand hiding in jean pockets. Only the noise of crumpled paper can be heard and, when this fades away, the room is deathly silent. A thousand pupils have been gathered here for six hours to answer a difficult question. All interaction with their peers is forbidden and, if an unexpected memory lapse should halt their train of thought, they cannot consult their notes. The essays produced by the pupils will sink into oblivion, stored in a dedicated hanger that has housed examination papers for many generations.
A few years later, I’m facilitating an all-day workshop in a large white room with some twenty computers. Around me, groups of pupils talk, laugh and see-saw between a sheet of drawing paper and the computer screen. Some isolate themselves to code, others are hunched over a 3D printer producing an open source design that they’ve just downloaded. The pupils consult their teachers, ask advice from the experts present in the room and share their progress with the others. Some momentarily leave their own group to help their friends in a competing group. The workshop involves “remixing” artistic works that have come into the public domain or are open source. No assessment is planned; the reaction of those present is the only measure of the quality of their production. Watching them, I think that they are extremely lucky to be able to draw freely from all these wells of existing knowledge: their own intelligence, that of their peers and teachers, virtually everything that humanity has produced and, above all, the global knowledge which is within easy reach. At the end of the workshop, we find their work surprising and original and the quality exceeds all our expectations. Our doubts about the pupils’ capacity to open up the raw materials and extract a structured form from them were unfounded. Now they make us smile.
We are connected to an infinite number of individuals, organisations and machines. In my view, the cooperation of all of these entities, regardless of the nature of their intelligence, is what defines collective intelligence
I observe the magic of collective creation every day at OuiShare, a collective project working to develop the collaborative economy. The project brings together people from all four corners of the world and I’m very lucky to be involved. Every day, with every project that we conduct, with every decision that we take and each disagreement that arises, we are experiencing intelligent cooperation. Within these laboratories for ideas and practices, we are committed to supporting the collaborative projects that spring up in the kitchen, in co-working spaces, during meetings. We are also endeavouring to learn, within our community, how we can create better together than each of us alone.
This is the alchemy of collective intelligence. Through cooperating together, we create and think better than alone, secluded monk-like in the monastery of our brain. We now have immediate access to the huge amount of existing knowledge but it is with others, today and tomorrow, that we create well. We are connected to an infinite number of individuals, organisations and machines. In my view, the cooperation of all of these entities, regardless of the nature of their intelligence, is what defines collective intelligence.
Will our interactions lead us to improve as individuals, as a species, or will they seal a new era of digital war?
For tomorrow’s world, the stakes are high with regard to the evolution of our thinking-together and deciding-together. We also have new companions which constantly assist - machines, programmes, robots – and which modify our ways of acting and thinking as much as we fashion them. These upheavals in our existence and our organisational models are today gathering such pace that the questioning of processes and of the effects of these interactions is acquiring a previously unseen consistency. We can no longer ignore the fact that we humans will no longer ever be alone.
In this critical context, how are we to define collective intelligence and to integrate machines into the production of future knowledge? Will our interactions lead us to improve as individuals, as a species, or will they seal a new era of digital war? If we want to consciously use our capacity to cooperate to make the world a better place, which economic, social, ethical and technological models should we build?
From the noosphere to collective intelligence
The telos of collective intelligence is part of the concept of the noosphere, created by Vladimir Vernadsky and analysed at length by Teilhard de Chardin. Understood to be the entirety of human thought, the noosphere corresponds to two phenomena in reciprocal interaction. On the one hand, the growing complexity of human societies from a cultural, social, economic and demographic point of view is moving toward the constitution of a sphere of ever-developing knowledge. On the other hand, this sphere, born from the multiplication of the ever-increasing number of interactions, is bringing about a gradual structuring of global thought and awareness by humanity of itself. The idea of a march towards some sort of human brain that transcends us, as old as it is[1], is particularly strong at a time when 40% of the planet is connected to the web. So, collective intelligence can be understood as a knowledge-creation process that is informed by the awareness of a noosphere.
The noosphere underpins the possibility of collective knowledge production, but it does not respond to the questions being asked if we examine the process of co-creation. A practical approach to collective intelligence enables us to explore the conditions favourable to the collective exercise of the intelligence of individuals, entities or machines.
Of networks and men
For this purpose, I would like to discuss the work carried out by the MIT Center for Collective Intelligence[2]. The research and analysis conducted by this Center are unique in their field. By combining mathematical, physical, biological, social and economic sciences and a resolutely forward-looking approach, the Center’s work aims to answer the following question: how can people and computers be connected so that — collectively — they act more intelligently than any individuals, groups, or computers have ever done before? The magnitude of the task does not intimidate Thomas Malone, founder and Director of the Center. In his opinion, the stakes are high for this research, because “Our future as a species may depend on our ability to use our global collective intelligence to make choices that are not just smart, but also wise.”[3] The practical significance of collective intelligence is beginning to take shape: on the one hand, it’s a matter of finding such a configuration that allows co-creation to result in ordered, efficient and useful choices that address certain ethics. On the other hand, is it reasonable to assume that a configuration that favours intelligent co-creation between individuals can also integrate machines?
As Thomas Malone aptly points out, collective decisions can absolutely be rational and stupid[4]! Consequently, the notion of intelligence must be expanded to integrate factors other than mere rationality. Thomas Malone defines it in this way: “to be intelligent, the collective behaviour of a group must exhibit characteristics such as perception, a capacity for learning, judgement and problem-solving ability.” In other words, the aptitudes of a group and those of individuals must function as communicating vessels: in a configuration conducive to co-production, the group thus acquires a range of behaviours that would normally be associated with a sole individual.
Collective intelligence thrives in organisational models that are structured in networks, distributed, decentralised and focused more on perception and listening than on rigid rules
The MIT Research Center then sought to determine the factors that correlate to more intelligent collective production. It seems that the average intelligence of each individual is not one of these. However, two factors significantly stand out: the degree of empathy among the group members and equal speech distribution within the group. Empathy, distribution and equality, these are factors that suggest that collective intelligence is at odds with hierarchical, compartmentalised and centralised organisational models. Conversely, collective intelligence thrives in organisational models that are structured in networks, distributed, decentralised and focused more on perception and listening than on rigid rules. It is not surprising that collaborative networks such as Wikipedia prosper: they present the exact characteristics that stimulate collective intelligence!
In my view, an extra ingredient is necessary to ensure that the multiplicity of the individuals making up the network does not result in stowaways. In this regard, we should remember that only 10% of Wikipedia readers are active contributors. The anonymity of contribution plays a role: the value produced by each contributor is neither measured nor recognised. Conversely, within Sensorica[5], an open network in which a group of individuals and organisations produce hardware solutions in a contributory manner, the added value of each contributor is regularly measured by the other contributors and acknowledged by the network. Peer evaluation and recognition of the value of everyone’s contribution are thus as important as the evaluation of the global value of the network. As Pierre Lévy wrote: “the foundation and purpose of collective intelligence are mutual recognition and individual enrichment rather than the culture of a fetishized and hypostasised community.”[6]
An intelligent network brings as much to the world as to its contributors: all for one and one for all. As a genuine place of learning, the network favours free circulation of knowledge and the exchange of decisions while respecting everyone’s contribution. Unlike organisational models in which the collective crushes the individual, an intelligent network is both the extension and the catalyst of everyone’s intelligence. An intent to collaborate and an awareness of the value thus created are essential to ensure that collective intelligence may operate.
Of machines and men: towards cooperation between complementary intelligences
Empathy, perception, judgement, awareness, intentionality: are these not actual human attributes? How can machines, in principle deprived of these attributes, be integrated into an intelligent network? Yet, above, when I evoked the implementation of a network of entities and individuals in order to determine the optimal organisation for the collective production of value, I wasn’t excluding machines. These are today widely accepted as the extension of human means and the idea of the forthcoming advent of singularity has a growing number of enthusiasts[7]. Today, the complexity and intelligence of computer programmes are such that we have reached a point of no return which, according to Kevin Kelly[8], arises when “technology alters us as much as we alter technology”.
In my opinion, the concept of machines as an assistant that is perfectly dominated by man is just as questionable as the faith in the superiority of machines’ intelligence over our own. On the one hand, computer programmes have calculative and data-analysis capacities that clearly exceed the capacities of human intelligence. On the other hand, the robots designed today are not only capable of duplicating themselves but also of learning and evolving[9]. The research undertaken by INRIA (French Public Institute for Research into Digital Sciences) focuses on cognitive development which is stimulated by curiosity, perception and representations.
When compared to the scale of human evolution, this progress has shown unprecedented speed. If the rate of progress of the past few years is maintained in the coming years, it is not fanciful to imagine that the robots of the future could understand and reproduce emotions, and self-generate programmes based on internal and external information in order to autonomously demonstrate thought, emotion and action. This autonomy, if it occurs, bestows machines with attributes which have until now been human characteristics: awareness, perception, autonomous production. Objectively, there are not sufficient scientific data to completely rule out autonomous technology today; it is therefore more sensible to assume that it is possible, whatever the timeframe may be. Conversely, technological evolution foreshadows a future in which man, not content with improving computer programmes, would be equipped with the technological means to enable an intervention on himself, a physical improvement or, why not, a behavioural (moral) improvement.
This vision very quickly starts to resemble a science-fiction scenario in which the machines, endowed with autonomy and awareness, end up rising against the human yoke in order to dominate us or to simply demand the same rights as our species. We are never far from the dialectics of master and slave: we cannot stop ourselves from transposing historical scenarios to tomorrow’s world. Behind this thought by analogy lies the visceral fear of being deprived of our means of control, since the machines that we design will be infinitely quicker and more efficient than us. The anxiety with regard to the coming ethical upheavals is often adorned in the finery of precautionary principle: since we are not absolutely certain that technology will present no danger for humanity, let’s slow down and, better still, let’s sound the death knell of its ambitions[10].
We stand at a historic moment where humans and technology are no longer two spheres capable of evolving without altering each other
Does this mean that we can hypothesize that technological progress is absolutely autonomous in relation to all ethical questions and, therefore, that the acknowledgement of the consequences of the humanisation of machines and of the sudden entrance of the mechanical world into that of the living has no place in a researcher’s laboratory? I don’t believe so, for the technologies that we produce are not artefacts and we cannot avoid the repercussions that they will have on the coming world.
In the face of these two stances – anti-technological and a-ethical – the hypothesis of cooperation between human intelligence and mechanical intelligence is, in the current stage of our knowledge, reasonable and desirable. We must however acknowledge that machines can deploy intelligence that is not only calculating and that, if it is to be different, will not necessarily be inferior to ours. There seems little doubt that this trend is causing upheavals that the human species has never experienced. However, to slow science down because we are struggling to realise the acceleration of technological progress would be an impasse. On the contrary, it’s up to us to imagine and implement the means of cooperation that fertilise the shared production of knowledge, learning and, above all, awareness. We stand at a historic moment where humans and technology are no longer two spheres capable of evolving without altering each other. Technology is as much our extension as we are its extension, for the future of our species henceforth depends as much on ecology as on technology.
I will conclude by saying that the new distributed organisations favour co-creation between men as much as they do between men and machines. The diversity of the entities making up the network, combined with the recognition of everyone at their fair value and depending on their means, represents a fertile land on which collective intelligence may flower.
[1] The noosphere, in one respect, is no other than the cosmos of Platonists, whereas the awareness of humanity has undeniable Socratic overtones.
[3] Interview with Thomas Malone, http://edge.org/conversation/collective-intelligence
[4] Ibid.
[6] Pierre Lévy, Collective intelligence: mankind's emerging world in cyberspace
[7] The notion of technological singularity was popularised by Vernor Vinge, a mathematician and science-fiction writer. In his essay The Coming Technological Singularity (1993), he attempted to demonstrate, using Moore’s Law, that humanity would have the technological means to produce superhuman intelligence within 30 years.
[8] Kevin Kelly, What Technology Wants (2010), Kindle version
[9] As shown by the Poppy project, an open-source robot designed by INRIA and capable of learning and adapting.
[10] For example, the Neo-Luddite movement. The most famous example of anti-technological activism is Ted Kaczynski, an American mathematician convinced that technological progress will lead humanity to apocalypse and that only the total eradication of modern society will prevent disaster. Known as the Unabomber, he sent sixteen bombs to famous figures in the technology world, killing three people and injuring twenty three.