“Hi,” siad the boy. “Want to play?”
The robot blinked. “Affirmative!”
from Boy + Bot
written by Amy Dyckman
illustrated by Dan Yaccarino
Alfred A. Knopf, 2012
(accessed on YouTube 8/27/23)
Sometimes I lose a word. I sometimes struggle to remember the name of someone when I meet them again. Not very often, I lose my train of thought. But these aren’t the types of language skills I’m thinking about.
Some people are non-verbal. They don’t have expressive language. Some of my daughter’s special ed students are non-verbal. They use a picture board to point to how they are feeling, what they need, or answer a question she poses.
People cannot speak for many reasons. But scientists are experimenting with ways to help people who have not lost their language skills, but are hampered physically because of paralysis, loss or weakness of the muscles used to talk, or anomalies in their voice box.
We’re familiar with Stephen Hawking’s computer-generated speech. Due to a tracheostomy performed to save his life after contracting pneumonia, Dr. Hawking was provided with a computer linked to a speech devise. It was adapted to allow him to use his thumbs to move a cursor and painstakingly point, letter by letter, to produce words that the computer voice spoke for him at about 15 words per minute. English conversations in the United States average about 150 words per minute. As Dr. Hawking’s ALS progressed, he was unable to use his thumbs. His devise was adapted again to respond to only one cheek muscle.
Since the 1990s neurologists and computer scientists have been developing ways to use AI to convert a persons’s thoughts into an action. BCIs (Brain-Computer Interfaces) use electrical signals in the brain, generated by a person’s thought waves, to accomplish an action. For example, a person with a prosthetic arm or leg or finger can imagine moving that limb or digit and through the BCI, it will move.
Says Seong-Whan Lee, professor of artificial intelligence and brain engineering at Korea University in South Korea, “We are now extremely interested in generating voices…from imagined speech.” Lee continues, “The main aim of our research is using [imagined speech] to analyze what a person is thinking and predict what they want to say…as these can be very direct and intuitive.”
Think of the message app on your phone filling in words before you finish typing them. Its choices are based on words and patterns you’ve used before and allows you to type a sentence word by word or even phrase by phrase instead of letter by letter.
Adding a voice component will enable a person who is unable to speak to become part of a conversation.
Applications for the technology are many and far-reaching. New England Journal of Medicine reported recently that a “[m]an who is unable to move or speak can now generate words and sentences on a computer using only his thoughts.” It’s due to an experimental brain implant that decodes signals once sent from his vocal tract, but is now unable to because of his paralysis. Work is also being done at Emory University and Georgia Tech.
Called neuroprosthesis, the device could help people who have suffered a stroke, survived a traumatic brain injury, or are living with ALS or other degenerative, paralyzing diseases.
Science has given us breakthrough after breakthrough regarding our complex brains. The biggest hurdle is not being able to decode brainwaves into words or even generate understandable, natural-sounding language.
The biggest challenge will be to provide people who will benefit from this brand-new technology the privacy necessary for them to keep some thoughts to themselves. So far, a neuroprosthesis does not come equipped with a door to close and lock against intruding eavesdroppers.
AI in all its many, many iterations, from autonomous vehicles to ChatGPT, to severe weather alerts depends on the ability of a computer or a robot controlled by a computer to solve problems by learning from its mistakes.
And it’s coming, ready or not. Some of it is scary. It’s (almost?) impossible to tell a fact from a fallacy, an original photograph from one that’s been photo-shopped, or a human from a hologram.
And while some of it is practical and useful like weather predictions, and all the “smart” appliances, devices, and cars, some of that is scary, too.
I’m trying to enjoy The Measure by Nikki Erlick. While I understand the importance celebrating the mundane, trying to transform regrets into positive action, and the threat of how prejudice affects society, I’m distracted by the premise: the strings and boxes themselves. And while I have a sense of my own mortality, I’m not sure I’d be any better (or worse) off knowing when my end will come. Or when anyone else’s will. But I’ll keep reading.
-—be curious! (and forge ahead with caution)
The robot blinked. “Affirmative!”
from Boy + Bot
written by Amy Dyckman
illustrated by Dan Yaccarino
Alfred A. Knopf, 2012
(accessed on YouTube 8/27/23)
Sometimes I lose a word. I sometimes struggle to remember the name of someone when I meet them again. Not very often, I lose my train of thought. But these aren’t the types of language skills I’m thinking about.
Some people are non-verbal. They don’t have expressive language. Some of my daughter’s special ed students are non-verbal. They use a picture board to point to how they are feeling, what they need, or answer a question she poses.
People cannot speak for many reasons. But scientists are experimenting with ways to help people who have not lost their language skills, but are hampered physically because of paralysis, loss or weakness of the muscles used to talk, or anomalies in their voice box.
We’re familiar with Stephen Hawking’s computer-generated speech. Due to a tracheostomy performed to save his life after contracting pneumonia, Dr. Hawking was provided with a computer linked to a speech devise. It was adapted to allow him to use his thumbs to move a cursor and painstakingly point, letter by letter, to produce words that the computer voice spoke for him at about 15 words per minute. English conversations in the United States average about 150 words per minute. As Dr. Hawking’s ALS progressed, he was unable to use his thumbs. His devise was adapted again to respond to only one cheek muscle.
Since the 1990s neurologists and computer scientists have been developing ways to use AI to convert a persons’s thoughts into an action. BCIs (Brain-Computer Interfaces) use electrical signals in the brain, generated by a person’s thought waves, to accomplish an action. For example, a person with a prosthetic arm or leg or finger can imagine moving that limb or digit and through the BCI, it will move.
Says Seong-Whan Lee, professor of artificial intelligence and brain engineering at Korea University in South Korea, “We are now extremely interested in generating voices…from imagined speech.” Lee continues, “The main aim of our research is using [imagined speech] to analyze what a person is thinking and predict what they want to say…as these can be very direct and intuitive.”
Think of the message app on your phone filling in words before you finish typing them. Its choices are based on words and patterns you’ve used before and allows you to type a sentence word by word or even phrase by phrase instead of letter by letter.
Adding a voice component will enable a person who is unable to speak to become part of a conversation.
Applications for the technology are many and far-reaching. New England Journal of Medicine reported recently that a “[m]an who is unable to move or speak can now generate words and sentences on a computer using only his thoughts.” It’s due to an experimental brain implant that decodes signals once sent from his vocal tract, but is now unable to because of his paralysis. Work is also being done at Emory University and Georgia Tech.
Called neuroprosthesis, the device could help people who have suffered a stroke, survived a traumatic brain injury, or are living with ALS or other degenerative, paralyzing diseases.
Science has given us breakthrough after breakthrough regarding our complex brains. The biggest hurdle is not being able to decode brainwaves into words or even generate understandable, natural-sounding language.
The biggest challenge will be to provide people who will benefit from this brand-new technology the privacy necessary for them to keep some thoughts to themselves. So far, a neuroprosthesis does not come equipped with a door to close and lock against intruding eavesdroppers.
AI in all its many, many iterations, from autonomous vehicles to ChatGPT, to severe weather alerts depends on the ability of a computer or a robot controlled by a computer to solve problems by learning from its mistakes.
And it’s coming, ready or not. Some of it is scary. It’s (almost?) impossible to tell a fact from a fallacy, an original photograph from one that’s been photo-shopped, or a human from a hologram.
And while some of it is practical and useful like weather predictions, and all the “smart” appliances, devices, and cars, some of that is scary, too.
I’m trying to enjoy The Measure by Nikki Erlick. While I understand the importance celebrating the mundane, trying to transform regrets into positive action, and the threat of how prejudice affects society, I’m distracted by the premise: the strings and boxes themselves. And while I have a sense of my own mortality, I’m not sure I’d be any better (or worse) off knowing when my end will come. Or when anyone else’s will. But I’ll keep reading.
-—be curious! (and forge ahead with caution)
RSS Feed