Can human beings, with the help of smart machines, not merely avoid “collective idiocy” (in Sandy Pentland’s words), but actually achieve a degree of intelligence previously unattainable by either humans or machines alone? These three panelists study the possibilities from different angles.
Thomas Malone’s Center for Collective Intelligence examines such evolving intelligent systems as Wikipedia, which relies on a veritable army of volunteers to “create a high quality product with almost no centralized control,” and Google, with its technology “harvesting knowledge” and serving up answers to a vast audience of seekers. While a crowd doesn’t guarantee the best solution to a problem, Malone sees opportunities in “prediction markets,” where humans, with the computational help of computers, predict things with greater accuracy than single experts, whether in electoral politics, or in medical diagnostics. Malone’s research is also attempting to set up metrics to measure the intelligence of these new human group-machine hybrids, and ways of applying collective decision making to climate change policy.
Alex (Sandy) Pentland performed a unique experiment in a large German bank, tagging its employees with special badges that tracked individuals’ interactions, down to head nodding, body language, and tone of voice. His research, conducted over a month, looked at how face to face interactions played into the overall organizational flow. The patterns he uncovered in the data collected from his name badges and from email and more traditional documentation, demonstrated the significance of social dynamics in workplace productivity. Certain individuals acted as information bottlenecks; others as polarizers, group thinkers, or gossip mongers. Pentland shared information about these patterns of communication with individuals. “Rather than think of this as big brother,” says Pentland, “think of this as a personal intelligence tool that collectively produces better results.” Related technology might be able to detect depression by examining a person’s patterns of socialization.
Karim R. Lakhani says he “stumbled into collective intelligence and distributed information systems as a puzzle.” While trying to market his large corporation’s medical imaging system, he discovered that a small Canadian group had “leapfrogged” his R&D team. A community of radiologists and physicists pooled their expertise to improve imaging technology, and beat a large, centralized lab. Since that time, Lakhani has pursued other examples of decentralized groups of people with a wide range of motivations, efficiently cracking complex problems-- from the open source software community, to biotech labs and entrepreneurial ventures. A T-shirt company, Threadless, asks its online community of a half million to submit T-shirt designs, and vote on them. The best scoring designs go into production. Sales are closing in on 1.5 million shirts at $20 a pop. Says Lakhani, “One hope of collective intelligence is that it takes the distributed and sticky pockets of knowledge that exist in the world and finds ways to aggregate them for us.”
Thomas Malone’s Center for Collective Intelligence examines such evolving intelligent systems as Wikipedia, which relies on a veritable army of volunteers to “create a high quality product with almost no centralized control,” and Google, with its technology “harvesting knowledge” and serving up answers to a vast audience of seekers. While a crowd doesn’t guarantee the best solution to a problem, Malone sees opportunities in “prediction markets,” where humans, with the computational help of computers, predict things with greater accuracy than single experts, whether in electoral politics, or in medical diagnostics. Malone’s research is also attempting to set up metrics to measure the intelligence of these new human group-machine hybrids, and ways of applying collective decision making to climate change policy.
Alex (Sandy) Pentland performed a unique experiment in a large German bank, tagging its employees with special badges that tracked individuals’ interactions, down to head nodding, body language, and tone of voice. His research, conducted over a month, looked at how face to face interactions played into the overall organizational flow. The patterns he uncovered in the data collected from his name badges and from email and more traditional documentation, demonstrated the significance of social dynamics in workplace productivity. Certain individuals acted as information bottlenecks; others as polarizers, group thinkers, or gossip mongers. Pentland shared information about these patterns of communication with individuals. “Rather than think of this as big brother,” says Pentland, “think of this as a personal intelligence tool that collectively produces better results.” Related technology might be able to detect depression by examining a person’s patterns of socialization.
Karim R. Lakhani says he “stumbled into collective intelligence and distributed information systems as a puzzle.” While trying to market his large corporation’s medical imaging system, he discovered that a small Canadian group had “leapfrogged” his R&D team. A community of radiologists and physicists pooled their expertise to improve imaging technology, and beat a large, centralized lab. Since that time, Lakhani has pursued other examples of decentralized groups of people with a wide range of motivations, efficiently cracking complex problems-- from the open source software community, to biotech labs and entrepreneurial ventures. A T-shirt company, Threadless, asks its online community of a half million to submit T-shirt designs, and vote on them. The best scoring designs go into production. Sales are closing in on 1.5 million shirts at $20 a pop. Says Lakhani, “One hope of collective intelligence is that it takes the distributed and sticky pockets of knowledge that exist in the world and finds ways to aggregate them for us.”
No comments:
Post a Comment