How can combining humans and high tech solve the world’s problems? How can we bring together the talents of both humans and computers?
We’re talking about human computation. Uniting a computer’s ability to analyze vast quantities of data with a human’s ability to think outside the box. A human-machine collaboration that could be used to solve the world’s most complex problems—from climate change to geopolitical conflict.
Does human computation already exist?
Yes. And you probably already know about it. Take Wikipedia: a free encyclopedia that comes in hundreds of languages and which functions through linking technology and internet users everywhere.
This is also known as microtasking: small tasks carried out by people all over the internet, together contributing to a single project. It’s like next-level multitasking. Instead of one single person writing entries for the encyclopedia, everyone writes them together, and everyone revises and edits them too.
Another example of human computation is Duolingo. This website and app offers free language lessons. While users learn the language, they’re simultaneously translating web content, which is how the app makes money. The combined efforts of all these language learners across the internet contributes to a single, professional-level translation.
Then there’s Foldit, an online puzzle video game about protein folding and also an experimental research project. Participants were asked to fold virtual proteins in the most efficient way possible. The aim was to discover how proteins fold so quickly and efficiently—a major question in molecular biology. The project discovered the tertiary structure of a regulatory protein for the pros-simian immunodeficiency virus, which could lead to new ways of treating AIDs.
A final example is Zooniverse, a citizen science web portal that promotes crowdsourced scientific research. In one Zooniverse project, known as Galaxy Zoo, participants are shown images of galaxies and asked to identify them. Other projects ask participants to identify craters on the moon or find planets around stars.
How can we take human computation to the next level?
In a recent editorial in the journal Science, researchers Pietro Michelucci, director at The Human Computation Institute, and Janis Dickinson, Director of Citizen Science at the Cornell Lab of Ornithology, envision a new and more powerful model of human computation. They call it a “dynamic Wikipedia.”
This is an online space where ideas could be shared, evaluated, and revised. The goal? To improve our understanding of current global issues and test out possible solutions. These solutions could then be implemented in the real world.
Think of it like SimCity, the video game series where gamers build cities from scratch. Essentially, it’s a model of the real world where people could try out different ideas, test out solutions, and explore the outcomes.
Tell me more…
Michelucci and his team have drawn out a framework that pushes the idea of Wikipedia to a new heights. Rather than simply a database of documents, they propose a simulation of the real world where people could try to solve real-world problems.
For example, an engineer might describe how an engine works, creating an online simulation of an engine. Then other participants might come along and pour in new fuel combinations and see what the outcome is.
Anyone could do this, from scientists to schoolchildren. Computer algorithms would then create a feedback loop that would constantly evaluate and revise the ideas within this shared space.
What could this human computation system be used for?
Examples suggested in the editorial include Project Houston. The idea is to use state-of-the-art speech analysis and natural language understanding to detect and offer help to depressed people or those considering suicide. Help would come in the form of the combined personalities of different individuals on the platform, using different expertise from different people.
Another area where human computation could help is in medicine. Machine algorithms can’t yet reliably identify tumors on X-rays, but humans are great at it. Participants could begin identifying easy images of tumors and then progress onto more difficult ones. It’s online learning, but also work.
But isn’t that AI?
It’s almost like artificial intelligence, yes. But AI is about trying to create intelligent machines that behave and react just like humans. Human computation, on the other hand, is about humans collaborating with machines, rather than creating machines to replace us.
We haven’t yet managed to create a computer as smart as a human. So advocates of human computation have decided to build humans into the computer system instead. In contrast to AI technology, machines are unlikely to become so smart that they pose a threat, because people will be inside the system to stop them. The technology remains in the hands of humans.
Although, of course, you might ask if humans can be trusted any more than machines? Human computation systems could still be abused; for example, being used to steal information or manipulate behavior.
Is human computation a good thing?
The emerging science of human computation definitely has a lot of potential. Through the interconnection of humans and machines, working together as a single system, we might be able to solve some of the most pressing concerns facing humanity today.
But only as long as the ethical questions can be addressed. How can a human-machine collaboration be designed so that it is meaningful? How can it protect and help the vulnerable? And, most significantly, who would fund these systems and therefore set their agenda? These are the issues that must be tackled head-on from the next generation of tech change-makers.