This title sounds like a badly written plotline for a science fiction film or a horrific nightmare for anyone with ill-thoughts, but it is not. Thanks to MIT’s AlterEgo headset, users can easily interface with devices through silent speech.
According to MIT’s Media Lab, AlterEgo is a “non-invasive, wearable, peripheral neural interface that allows humans to converse in natural language with machines, artificial intelligence assistants, services, and other people without any voice – without opening their mouth, and without externally observable movements – simply by articulating words internally.”
In other words, you can send words from your head just by thinking them. This goes beyond the usual virtual assistants like Alexa or Siri, which need to detect voice commands to function. This is the next evolution in this arena: gadgets that can listen to the voice inside your head.
To “listen” to your inner thoughts, the device picks up on the electrical signals produced by small movements of our facial and neck muscles when we read or talk to ourselves silently. The device itself is a 3D-printed plastic gadget extending from the back of the ear to below the bottom lip:
Figure 1. MIT Media Lab (Lorrie Lejeune/MIT)
Developed by Arnav Kapur at MIT, AlterEgo was programmed with a few simple applications like game-playing assistants for chess and Go to suggest the next move, as well as an app to answer math problems. Additionally, AlterEgo is 92% accurate in its ability to capture what users said.
Nonetheless, the technology still faces various limitations due to its nascent nature and the scarcity of large data sets capable of training AlterEgo’s algorithm, compared to a typical speech-recognition app. Furthermore, the person-to-person communication is limited to simple words and phrases, and the translation capabilities are modest.
The application for the technology are limitless. Kapur recently announced he is setting up a study at hospitals and rehabilitation centres where people with speech impairments will use AlterEgo.
Read more on the same subject: