Artificial intelligence (AI) has long held a place in pop-culture, but whilst it isn’t yet intelligent enough to mirror the robots we see in movies, the last few years have seen rapid advances in the field. This has been aided by increases in processing power and available storage. The tech industry is leading the way, with Facebook developing software which can describe images to visually impaired people, and Skype’s ability to translate from one language to another. However, there is still a lot of work to be done to overcome the communication barriers between AI systems and humans in order to allow us to communicate with them effectively. This month saw an important development in that direction, when the Computer Science and Artificial Intelligence Laboratory (CSAIL) at MIT published their work on Baxter, a robot which can be controlled by our thoughts.
Breaking down the learning and teaching barrier between robots and humans may sound like science fiction, but CSAIL have used signals from an electroencephalography (EEG) monitor to allow humans to correct Baxter’s mistakes in real-time. A person sits opposite Baxter with a cap of electrodes attached to their head. When they see his mistake, brain signals called ‘error-related potentials’ are generated and converted to code which is transmitted to Baxter, who realises his mistake and corrects it. Currently, the system is almost 80% accurate. A simple idea, with enormous potential.
Communication with robots prior to this research has been unnatural and slow. It involves typing or vocalising commands, often involving complex learning algorithms to try to ‘teach’ the robots how to act in a more human manner. If we want them to understand our language, we have to teach them, although most Scottish natives will be all too familiar with the failings of this approach. CSAIL’s research aims to allow us to communicate with robots in a more natural, human way. While Baxter can only understand whether he is correct or incorrect at the moment, the CSAIL team is looking to the future of the technology and its potential applications in daily life.
“Imagine being able to instantaneously tell a robot to do a certain action, without needing to type a command, push a button or even say a word,” says CSAIL Director Daniela Rus. “A streamlined approach like that would improve our abilities to supervise factory robots, driverless cars, and other technologies we haven’t even invented yet.”1
There are endless possibilities for this technology, not least those which it opens up for people who are unable to communicate verbally. As the technology expands, the opportunity for those with verbal impairments to use it to complete tasks seems incredibly exciting.
This article was written by Dawn Gillies and edited by Teodora Aldea.