21st Century Common Sense: Collective intelligence for the Climate Crisis
To create a learning network whose speed matches the pace of change in development, we need a new kind of intelligence. We’re working with Nesta’s Centre for Collective Intelligence to learn how to tap into distributed, real time intelligence as a global good.
To start, this means learning to see in a new way — getting more real-time data. This work has started in exciting prospects, but it is not yet operating at scale. We need more of it. While a minister may need to make development investment decisions with data that is two years old, who would cross the street based on data from 2017? Thanks to ground-breaking work at UN Global Pulse and others, we know that moving closer to real-time data is both possible and adds value, while managing risks related to data privacy.
But getting faster data is not enough. We also need to find better ways to channel people’s insights, views and activation towards intractable development problems. We’ll also need to learn to model new forms of accountability that tap into shifting and varying realities.
Nesta’s jargon-free approach to collective intelligence is a breath of fresh air. it’s about understanding the problem, finding solutions and making decisions about what to do. This needs to be done collectively because no one person (or algorithm) knows it all. Building on existing solutions in sustainable development, Nesta will help us codify collective intelligence as a way of seeing and operating in real time.
Time to see more crossovers: Machines, humans and wicked problems
Collective intelligence is people working together, sometimes with the help of machines to analyze issues, design solutions and inspire new forms of decision making. One way to think about it is that we are looking at a giant brain, a Big Mind, as Geoff Mulgan writes, where the functions of the brain — memory, analysis, creativity among others — are performed in a distributed way by many people together and often with the help of technology.
This work isn’t brand new. The United Nations has been advancing the use of collective intelligence:
- To predict displacement in Somalia, UNHCR is using machine learning through Project Jetson
- UN Global Pulse has used natural language processing to learn from people calling in to talk radio shows to track new agricultural threats and learn about people’s attitudes towards refugees and is leading on privacy and ethical frameworks to govern the use of big data in development.
- Here at UNDP, we have tested out ways of helping our Sudanese national partners gain real-time and detailed insights about poverty distribution using mobile phone metadata.
- Satellite data is also enabling us to develop a more precise understanding of the socio-political and geographic changes that are driving people to migrate.
- We have also seen crowdfunding evolve as an impact investment tool to finance startups in Egypt
These are a good start, but the field needs to advance if we are going to tap into collective intelligence to avert climate crisis. Within the Accelerator Lab Network, we hope to be a part of that. Together with Nesta, we’ll be looking for insights in these zones:
Learning direction #1: Moving beyond using collective intelligence to understand the problem
From what I see, the current development work framed as collective intelligence experiments focus on getting a better understanding of what the problem is. We tend to get stuck on that part. Much less attention is paid to collectively developing options and ideas to solve problems and an even smaller portion is testing out ways to drive more inclusive decision making on what should be done. A key area of learning for our partnership with Nesta is how best to get to solutions and public accountability using collective intelligence methods. The learning question at hand is: How does the hive mind get better at making decisions together and overseeing public investments?
Learning direction #2: Balancing public sector promises with the flexibility to follow the unexpected
A core principle of instilling trust in the public sector is that promises made are kept. Here’s where open data about where public money goes is a good basis, but it is insufficient. In a world of constantly changing wicked problems and new opportunities, sticking with the plan is often not enough.
Years ago, when working with a team on a crowdsourcing project in Eastern Europe, I was part of the effort to collect people’s views on city safety. We expected to hear about crime and violence, and instead learned that a core issue for people were packs of stray dogs who, once protected in gangs, were determining where people walked and where they felt threatened. Our mandate was not in this area, so we were stumped. Having heard people’s input, we couldn’t act. Our budget was locked into to what we and our partners had pre-defined as the problem.
This story may resonate with others who open up to learn from the crowd to define problems and solutions but then still need to operate within predefined budget tracks. So, a second learning question is: How does the public sector follow the unexpected results of collective intelligence without mission creep or wasting resources on unplanned activities?
Learning direction #3: Unite the system thinkers, the machine designers and those who study the human brain
Sustainable development is a real-time, distributed problem. The likelihood of future generations living at least as well as we do is dependent on billions of choices and actions every second: what crops we grow, what we build our houses with, how we get around, who we vote for, how we heat our homes and cook our food….It’s a daunting set of actions, choices and interactions. This is why we think this problem demands collective intelligence.
To collectively observe, remember and someday maybe even exercise the judgement needed for future-oriented action, we need to bring together three fields of work: those who study cognitive bias in humans, those who design machine intelligence, and those who understand complex adaptive systems. We need the the human experts, machine experts and system thinkers to work together if we are going to use collective intelligence to get the world on a more sustainable path.
Humans have their biases, whether acting as individuals or a collective. These biases make it tough to face climate change. David Wallace-Wells brings together behavioral psychology and our perception of climate change in the Uninhabitable Earth when he talks about multiple cognitive biases that make it hard to get people to act to mitigate the disastrous effects of climate change.
Some think that the digital age may be making it even harder to focus on complex problems. Douglas Rushkoff, for example, thinks it is getting more challenging for humans to do the systems thinking needed for climate change because our pinging phones are making us operate in fight or flight mode instead of thinking long term- which is the kind of thinking that is needed to overcome short term decisions that impact the planet. How does our networked information environment interact with cognitive biases, especially when it comes to acting in the face of climate crisis?
Machines could be one ingredient to help us work together like a giant multi-headed brain. But as we use machines to help us get smarter together to act responsibly in the face of climate crisis, we need to know more about the limitations of machine intelligence. As Molly Callahan of MIT writes “it’s time to study machines like we study humans.” We’ve seen how when machines learn from test data, this process can expand bias when it comes to predictive sentencing models and job applicant screening. While these are only early signals and in all fairness humans have demonstrated biases for much longer than machines, it’s still concerning. What if this type of phenomena continues when it comes to cognitive bias that discounts the future? If the machines amplify the bias that makes it hard to act against climate change in the way they have been known to amplify racial or gender bias, we’re really in trouble.
If we want to tap into what collective intelligence can do as a multiplier, we need to understand rather than amplify biases that makes it hard to take the future seriously. We need to embrace the serendipity of being open to multiple views and data points simultaneously. And we need the mindfulness on a global scale to be able to do more than to react; we need generate a new kind of intelligence and then proactively decide how to act to stave off climate catastrophe.
So if you are a climate change activist, a machine learning expert or someone working on nudges to overcome cognitive bias…. As the Beatles would say: Come together! Help us shape experiments towards a giant development brain that collectively senses what is out there and, together, decides what to do about it.
With thanks to Kathy Peach and Geoff Mulgan for input and inspiration, not least of which for the title of this blog! This blog is an attempt to iterate thinking from a talk at The Crowd earlier this year. Admittedly, the thinking might have been clearer in the talk. Watch it here.