AI Powered Hive-Mind Turns Groups of Networked Users into Powerful “Lie Detector”


Researchers from Unanimous AI and Oxford University presented a research study at the Future Technologies Conference 2017 demonstrating that groups of networked users, when connected in real-time systems online, can detect deceit in human facial expressions with significantly fewer errors than individuals can detect on their own.

The study — Using Swarm AI to Detect Deceit in Human Faces — engaged 168 randomly selected participants and asked them to view short video clips of smiling faces and assess if each video depicted an authentic smile induced by a true emotional response or a fake smile delivered upon command. Prior research shows that human subjects average a 33% error rate when discriminating smiles this way. Subjects in this new study confirmed the 33% error rate when assessing smiles on their own. But, when working together as AI powered “Hive Mind,” these same subjects were able to drop their misidentification rate to 18%, reducing errors by nearly half.

This research was performed using Swarm AI technology to assess deceit in the smile videos, a method that employs both human judgement and artificial intelligence algorithms to achieve results that are generally more accurate than either humans or software can do alone. The technology is based on the science of Swarm Intelligence, a biologically inspired process that goes back to the birds and the bees and other social creatures that amplify their group intelligence by forming flocks, schools, colonies, and swarms. Across countless species, nature show us that when working together as closed-loop systems, groups can produce collective insights that greatly exceed the intellectual abilities of individual members. While humans have not evolved this ability naturally, Swarm AI technology enables human groups to combine their knowledge, wisdom, insights, and instinct to generate optimized insights.

For context, here is the Smile referenced above….

The most well-known example of using Swarm AI technology to amplify the intelligence of human groups happened last year when Unanimous AI was challenged by CBS Interactive to predict the Kentucky Derby – not just the winner but the first four finishers in exact order. In horseracing, this longshot wager is called the Superfecta and last year it went off at 542-to-1 odds. Unanimous made the prediction by tapping the intelligence of only 20 horse racing fans, their insights combined in a Swarm AI. The prediction was perfect. As reported by Newsweek, anyone who put $20 on the Superfecta prediction won $11,800. One lucky reader won $50K.

“Over the last few years, many studies have shown that Swarm AI technology enables human groups to amplify their intelligence and make more accurate predictions,” said Dr. Louis Rosenberg, founder of Unanimous AI. “But this is the first study to demonstrate conclusively that artificial human swarms can also amplify social intelligence, enabling significant accuracy benefits in very human tasks.”

Want to learn more about our Swarm AI technology? Check out our TED talk below…

Every week Unanimous A. I. brings together swarms of regular people just like you to help us predict sporting events, investment opportunities, and technology as well as provide insight into everything from pop culture to politics. If you’d like information on joining a swarm – or just want to stay up to date on the latest predictions – click the button below.

Yes, Send Me The Newsletter