This is a piece just published by Project Syndicate
'One might not know it, judging by all the doom and gloom in the press, but there are still parts of the world where technology is regarded as a force for good – even for salvation. A recent survey shows that over 80% of young Africans are optimistic about what technology will do for the continent.
But elsewhere, people increasingly feel as though they are locked in a titanic struggle against technologies that threaten to take their jobs, steal their data, destroy the very idea of childhood, and disrupt democracy. Technology often seems like something that is done to us by distant and unaccountable forces, rather than something that we control. It doesn’t help that, for all of the extraordinary hype about artificial intelligence (AI), much of the investment in this domain has focused on military applications and manipulative ways to target propaganda and advertisements. No wonder people feel vulnerable and anxious.
Fortunately, there are still ways for us to develop a better relationship with the extraordinary technologies that are coming online and to the market. One alternative strategy is to develop “collective intelligence” (CI), which, rather than seeking ways to replace people with AI, focuses on combining the best of humans with the best of machines. This approach is becoming increasingly influential in business, science, and government, partly because it works, but also because it embodies the democratic, humanistic values that many of us hold dear.
Anyone who has used Wikipedia will have a sense of what collective intelligence means. Since the 1990s it has allowed millions of people to collaborate together online to share the world’s knowledge.
In some respects this isn’t a new idea. The Oxford English Dictionary recruited tens of thousands of volunteers in the 19th century to map the shifting meanings of English words, using methods quite similar to Wikipedia. But the field is taking off now mainly because the tools at our disposal are so much more powerful than ever before.
Take science. ‘Citizen science’ projects like Zooniverse now mobilise millions of people online to find new stars in the galaxy, observe nature or analyse tumours. The advocates of citizen science spotted that there is a huge surplus capacity of brain power – particularly in highly educated countries – which could be tapped into.
Healthcare is also being transformed by collective intelligence, with thousands of projects bringing patients together to share data or devise better ways of managing diseases. One famous project, prompted by a patient’s frustration with the medical establishment, mobilised volunteers to design an artificial pancreas (the ‘Do-It-Yourself Pancreas System’).
Business is also becoming engaged, though again with far less attention than AI. Duolingo uses volunteers to improve its service providing language teaching, adding in new language pairs. Lego has long used tens of thousands of keen consumers to help it design new products. Siemens uses collective intelligence methods to organise how it allocates funds internally – on the principle that its engineers will have a better idea of what’s likely to succeed than management.
NASA is a particularly striking example. Our image from the 1960s was of hundreds of men in white coats sitting in a room in Houston. NASA still employs many scientists. But it has opened itself inside out so that it can draw on ideas from anyone anywhere, whether to design a new spacesuit or programme software for rockets, offering financial rewards for the ones it uses.
The unifying idea here is that any organisation is more likely to succeed if it can find ways of mobilising data, insights and ideas from as wide a range of sources as possible, rather than just relying on the people inside its walls.
Many of the best collective intelligence examples now combine human brainpower and the processing power of computers rather than seeing them as alternatives. Chess showed the way. Many years ago it turned out that ‘freestyle chess’ – groups of people working with the help of computers - could often beat both the best individuals and the best computers.
This idea of combining AI and CI is now being applied in many fields. In one current project, based in Swansea University, people in Yemen upload images of munitions left on the ground so that machine learning algorithms can then classify them and help build legal cases around war crimes. A very different example is the city of Jakarta which combines data on flooding from citizens with data from sensors to create a real time observatory. Meanwhile many are experimenting with similar ideas – combining multiple sources of data - to better track and predict what could happen with Coronavirus.
Democracy is a particularly promising field for collective intelligence, with many experiments showing how technology can be used to open up power as an alternative to authoritarianism. Taiwan is a trailblazer, involving millions of citizens in shaping policy, helped by AI algorithms that show the state and shape of opinion.
Perhaps it’s not surprising that the UN – through its new ‘Accelerator Labs’ - has picked on collective intelligence as one of the keys to speeding up achievement of the sustainable development goals.
Hundreds of billions of dollars are being spent this year on AI, which works very well for some tasks – recognising faces, making recommendations as to what Youtube video you might want to watch or winning games of Go. But even these kinds of AI in practice depend heavily on people to train the algorithms. And for most of the complex, messy issues that matter most in daily life, AI on its own simply doesn’t work very well. For these we need to combine the machine and the human, the AI and the CI.
History tells us that purely technological fixes tend to be overhyped and tend to disappoint. The next decade will hopefully be one when we learn better how to use technology as much to enhance our own abilities as to replace them.
The PS piece can be found here:
