Millions of people have speech disorders from conditions like amyotrophic lateral sclerosis and multiple sclerosis. How can they communicate without actually talking? MIT media lab designed AlterEgo, a headset that enables users to communicate in natural language without opening their mouths or any obvious movements. When a user formulates a sentence in mind, such as a Google query, and articulates it internally, the headset’s sensors capture the neural signals sent from brain to the back of tongue or plate. The neural signals are processed by AI and transmitted over the Internet to a computer so that the sentence is processed, such as by a Google search. The result is returned to the headset’s speaker as a voice. Since the voice is heard through bone conduction, it does not disrupt the usual auditory perception. The system understands its user 92% of the time. This technology may have uses for the general population as a human-computer interface that uses natural language. It does not read minds or invade privacy since it only functions when the user actively tries to speak. The headset does not block ears and therefore does not unplug a user from the environment. AlterEgo has the potential to connect a human to machines, AI, and the Internet and augment her with information and enhanced cognition.
Summary
Purpose
To allow people to talk without moving
Idea
Capture and decode the neural signals sent to the tongue and mouth
Further Possibilities
1. Return information through eyewear
2. Use AlterEgo for dictation
3. Use AlterEgo to control smart cars
4. Make AlterEgo invisible
5. Use AlterEgo and Eye-tracking to operate computers without keyboard
Questions
1. How might we combine different signals sent from a human?
2. How might AI capture all the information communicated by a human?
3. How might we enhance human cognition?
