For well over a century, researchers in the field of Collective Intelligence have shown that groups can outperform individuals when making decisions, predictions, and forecasts. The most common methods for harnessing the intelligence of groups treats the population as a “crowd” of independent agents that provide input in isolation in the form of polls, surveys, and market transactions. While such crowd-based methods can be effective, they are markedly different from how natural systems harness group intelligence. In the natural world, groups commonly form real-time closed-loop systems (i.e. “swarms”) that converge on solutions in synchrony. The present study compares the predictive ability of crowds and swarms when tapping the intelligence of human groups. More specifically, the present study tasked a crowd of 469 football fans and a swarm of 29 football fans in a challenge to predict 20 Prop Bets during the 2016 Super Bowl. Results revealed that the crowd, although 16 times larger in size, was significantly less accurate (at 47% correct) than the swarm (at 68% correct). Further, the swarm outperformed 98% of the individuals in the full study. These results suggest that swarming, with closed-loop feedback, is potentially a more effective method for tapping the insights of groups than traditional polling.
In the natural world, many species amplify the accuracy of their decision-making abilities by working together real-time closed-loop systems that converge on optimal solutions in synchrony. Known as Swarm Intelligence (SI), the process has been deeply studied in schools of fish, flocks of birds, and swarms of bees. The present study looks at the ability of human groups to make decisions as an Artificial Swarm Intelligence (ASI) by forming similar real-time closed-loop systems online. More specifically, the present study tasked groups of typical sports fans with predicting English Premier League matches over a period of five consecutive weeks by working together in real-time as swarm-based systems. Results showed that individuals, who averaged 55% accuracy when predicting games alone, were able to amplify their accuracy to 72% when predicting together as real-time swarms. This corresponds to 131% amplification in predictive accuracy across five consecutive weeks (50 games).
In this study, users looked at videos of people smiling. The people in the videos were either smiling in response to something — real smiles — funny or simply because they were told to — fake smiles. Then the 168 test subjects were asked to evaluate whether the smiles were real.
The company founded by Louis Rosenberg has been extremely successful in creating a fun and scalable tool to augment group decision making. The company first tested their superpowers with UNU that was then rebranded as Swarm.ai for commercial purposes. But how can Swarm intelligence be uncannily so smart? Just luck? Perhaps. Or maybe Swarm intelligence taps into a more fundamental human capacity: uncertainty reduction.
Although simply pooling together judgments can give you pretty good solutions, this is often not enough when tasks are complex, judges have overlapping information and ground truths are unknown or uncertain. In these situations, the best strategy to adopt is confidence weighting. Multiple agents in group can optimally integrate information by not only sharing their preferences but also how confident they feel about their preferences .
Yes, overconfident individuals tend to screw up the group.
This trick (Bayesian information integration for the geeks) is also used by the brain when trying to integrate information from multiple modalities (e.g., the sound of a bird from your ears with the image of the bird from your eyes). And if I am reading your mind correctly, yes, overconfident people tend to screw up the group.
But confidence sharing does not need to be conveyed by language only. Actually, language is not ideal because it is not scalable (have you tried reaching a consensus in a room with more than 10 people?). Many more "confidence" signals can be and are indeed used. Voice intonation, posture, gestures, stubborness, all convey the same message "I am right, you are wrong". Some apes try to convince the group to move toward one direction by moving first and just sitting there. Interestingly, we now have the technology to use these principles online. What are we going to use it for?
E pluribus unum
Unanimous AI is a technology company that amplifies the intelligence of human groups using AI algorithms modeled after swarms in nature. Inspired by the intelligence amplification effects that occur within flocks of birds, schools of fish, and swarms of bees, Unanimous enables people to achieve similar benefits by forming "artificial swarms" online. Known as Artificial Swarm Intelligence (ASI), the core technology enables groups to efficiently combine their knowledge, wisdom, insights, and intuitions into an emergent intelligence that is sometimes referred to as a "hive mind."
Unanimous has deployed swarm-based technologies through a cloud-based server Swarm.ai, (formally UNU) which enables online groups to answer questions, reach decisions, and make predictions by thinking together as a unified intelligence. This process has been shown to produce significantly improved decisions, predictions, estimations, and forecasts, as demonstrated when predicting major events such as the Kentucky Derby, the Oscars, the Stanley Cup, Presidential Elections, and the World Series.
Unanimous is also the provider of Swarm Insight and Swarm IQ, two on-demand intelligence services using their SaaS platform called Swarm. In 2018, Swarm AI technology from Unanimous won the Best AI and Machine Learning Technology Award at the South by Southwest (SXSW 2018) Innovation Awards, as well as winning the overall Best in Show Award for technology at SXSW 2018. In 2018, Unanimous AI also made headlines with Stanford University School of Medicine by amplifying the intelligence of medical doctors using the Swarm platform, enabling significantly more accurate diagnoses than traditional methods or deep learning.