MIT is experimenting with a device which can read the words which you sub vocalise. The device is fixed to the user's face and uses electrodes to read electronic signals associated with "speaking inside your head" and AI to interpret them.
So far so good, this might have a number of practical applications and in an office environment would be better than speaking out loud to an intelligent agent such as Alexa, which would be intrusive and potentially annoying to other office occupants.
It also might have useful applications for communicating with people with speech disabilities. It might for example have sped up Steven Hawkin's communications via speech chips. The potential may exist for communicating with people with Locked in Syndrome etc.
However there is the potential for use in police interrogation and the technical ability over ride "the right to silence" or "not to self incriminate oneself", which exists constitutionally in certain judicial domains. This potentially provides a quite dangerous tool to state authorities in certain draconian regimes. Although, it could be considered more humane than water boarding.
As ever, with technological progress comes ethical questions about control of acceptable use of potentially beneficial innovations. So who is going to do this for the Direct Human AI interface?
No comments:
Post a Comment