A way to “Detect Speech” from People’s brain
We don’t just make any old random noise when we talk; we’re thinking about our words, and that makes us able to speak fluently.
Meta’s new AI can scan a person’s brainwaves to “hear” what someone else is saying to them. In other words, it can tell which words you hear by reading your brainwaves. This is not the first time this concept has gotten spotlight. In 2019, American scientists had developed an artificial intelligence that could accurately read brain signals and translate them into speech.
Talking about the Meta’s recent AI, it can decode speech from noninvasive recordings of brain activity. Neuroscientists have always dreamt of decoding speech from someone’s brain for a long time, but invasive methods were needed to achieve this.
The specialty of the new technique, according to the researchers, is that it is non-invasive, which means that researchers do not have to implant like… electrodes, in anyone’s brain.
Noninvasive techniques such as electroencephalograms, EEG, and magnetoencephalography, MEG, can scan the brain from the outside and watch activity without any surgery, but the problem is that they are too noisy.
In order to address this problem, researchers turned to machine learning algorithms to help “clean up” the noise. They used the model wave2vec 2.0.
Brain-wave-reading AIs seem to be an exciting new technology that can be used to help people with speech problems, like people who can’t speak, and those who have had strokes or other issues that cause speech difficulties.
But so far, only lab-based research has been done on brainwave-reading AI. They haven’t been available for use in the real world yet.
The future of Neural AI technologies
The future of AIs is will not be limited to be acting as a cure for people with speech problems. With advancements in technology, it’s not farfetched to think that advanced form of AI could be used as ways for computers to communicate with each other.
When we think about the future of technologies like Meta’s Brain-wave-reading AIs, we shouldn’t forget about the potential for misuse, should we?
In the future, will these kinds of hacks also be possible against our brains? Could criminals hack into our minds to get information from us?
Well, we do have trouble with people who spread false news stories on social media to stir up anger. But still, that’s too much for assumption.
In fact, it’s too childish to stop the progress of something new just because it has some potential for misuse. Let’s keep in mind that people can also use every single piece of technology for evil purposes, and they still do it.
Meta’s new steps are looking promising, and we are in the correct path among few, at least, till now.