We are one step closer to a world of robots and artificial intelligence as researchers have just released a new system that allows for humans to correct the mistakes of robots using only their brains.
Humanity has been making leaps and bounds in terms of artificial intelligence. We see it in everyday technology, from Facebook’s facial recognition system for tagging photos of certain people on their social media site all the way to servers and applications learning security functionality based off prior attacks. These systems are put in place to make human-like choices, where we use collected data in order to pick the best route moving forward. However, they can still be prone to errors and are far from perfect.
Correcting a failure like this can be a daunting and highly complex process. In order to avoid this pitfall, researchers from Boston University have set out to change that with their new artificial intelligence correction system.
The research team has revealed that artificial intelligence mistakes can be corrected in real time using brainwaves from a human that is observing the mistakes that a robot makes. Through the usage of electroencephalopgraphy, we are able to monitor brain activity and pass that information on.
One sample of this technology was done with a robot that had been nicknamed “Baxter”. The robot had the task of sorting various object. A human participant watched the robot sort the objects, and would acknowledge when a mistake was made. The brain waves would be picked up and sent to the robot to inform it that a mistake had occurred so that it could correct itself.
This new technology could be a huge step forward for artificial intelligence and machine learning.