Ketty de Falco
CEO, Insights Division
The unsuspected power of Collective Intelligence: pushing the limits of forecasting
Artificial Intelligence only goes so far when it comes to the future. Combining it with collective human intelligence can create a new paradigm in forecasting.
By 2020, we will live with seven connected devices per person, on average. It is estimated that 212 billion devices are compatible with sensors. Imagine the amount of data this constitutes. The exploitation of this information is at the heart of one of the most profound changes in the way we think about services, objects and brands.
The era of data and artificial intelligence
Today, declarative data from studies and surveys are no longer the main source of information about consumers. With the development of digital platforms and the explosion of data, there are now much richer banks, thanks to analytics, social media and many other sources to combine these insights with. Ultra-connected, the customer experience is also much more complex than in the past.
Artificial Intelligence appears to reduce this complexity and predict the future with more certainty. However, it is still difficult to understand the vastness of the ocean of behavioral data provided by these devices, because we still do not know its limits. Moreover, in the hyper-solicitation of data we have overly accelerated processing of information. Too often, we react to a shower of data with impulses, ignoring gaps in unverified information.
At Kantar we see a role for AI, but we also believe it cannot replace human intelligence. We see the future in integrating data with collective intelligence.
For the first time a few years ago, researchers wanted to measure the intelligence of a group’s collective intelligence, in the same way that we measure an individual’s IQ.
Two important discoveries emerged from their research:
- The measurement of a group’s intelligence is possible, and this number is predictive of the group’s ability to solve problems
- The intelligence of a group is greater than the sum of the individual intelligences and does not depend on individual IQ’s
What we have observed in nature for thousands of years has been proven on a human scale. Birds, fish, and bees have been able to amplify their intelligence by joining forces, thinking together in a system.
The collective intelligence they demonstrate reveals that, while numbers are a strength, they are not everything. The difference results from the density of connections and the organization of networks.
The principle of “wisdom of the crowds”
The reliability of the results obtained through collective intelligence is based on the principle of crowd wisdom: the average of the estimates is more reliable than an average individual estimate.
This phenomenon is now well documented in many areas. It is based on three principles.
- The average of the estimates is more reliable than an average individual estimate. The collective estimation of a group, whatever its purpose—the weight of an ox, a sports prognosis, an economic forecast, et cetera—is a collective estimate. It is generally more reliable than that of a member of a randomly selected group.
- The larger the crowd, the more reliable this estimate is. The power of the average is based on the mutual neutralization of individual errors. As a result, the larger the crowd, the more likely it is that one person’s mistake will be compensated for.
- The increase in reliability provided by each additional estimate decreases.
The diversity theorem
This theorem has been scientifically proven by Anita Woolley of Carnegie Mellon University, and Tom Malone, head of the MIT Center for Collective Intelligence.
All these analyses show that the most essential component of collective IQ is not the sum of individual IQs, but the emotional intelligence of the group members, as expressed in their ability to listen, empathize and respect the contributions of others. The quality of communication between brains is more important than the power of the brains themselves.
To obtain the best grouped estimate, it is necessary to use experts AND people who do not think alike. Expertise and diversity are the two essential components.
Recent analysis has found that that companies experimenting with big data increasingly find it difficult to make sense of. More and more companies are thus returning to historical research companies, which are gradually realizing their potential as data integrators. Data (especially declarative data) remains a valuable source of information, especially when it is linked to other types of data to deliver even more added value. At Kantar, we strongly believe in expanding knowledge through collective intelligence to push the boundaries of forecasting. We focus on using technologies, data and human beings’ collective intelligence, combining the power of machines with human creativity, the power of emotions, and human intuition.