Can AI Provide Communication Assistance to People with Aphasia?
Living with aphasia, a condition that affects one’s ability to understand and express language, can be incredibly challenging. Recent advancements in artificial intelligence (AI) have opened up new possibilities for empowering aphasia patients and transforming their ability to communicate. In this blog post, we explore how emerging AI technology could make an impact on the lives of individuals with aphasia.
New assistive technology
A new AI device called the semantic decoder was created by researchers at the University of Texas. The device is able to translate brain activity into written text. The technology is still in progress and not perfected. Currently, it is about 50% accurate.
As the semantic decoder is improved, it could be revolutionary for individuals experiencing aphasia for a few reasons:
1. It’s non-invasive
Similar versions of this technology require a surgically implanted device. However, the semantic decoder does not require surgery and is not invasive in any way. Instead, it works through fMRI scans measuring brain activity.
On the flip side, this makes it difficult to provide accurate results. The technology requires users to “train” it by listening to hours of podcasts while in the scanner. After the training is complete, the decoder is still currently accurate only about half of the time. The technology is not perfect at this time, but it’s an incredible step in the right direction.
Interestingly, this technology requires active participation from subjects, both to train the decoder and to use it. Therefore, it can only be used if the subject is willing, which is great news as it ensures mental privacy is respected and the decoder can’t be used in an immoral way.
2. It produces continuous text
Other similar decoding technologies provide subjects with a set list of words that they are able to use. Then, they can generally only produce short sentences based on these word lists. The semantic decoder is able to continuously produce text and learns new words based on the training it is provided.
Conclusion
While the semantic decoder is still being developed, it shows incredible progress for communication assistance for people with aphasia. Through its non-invasive usage and its ability to produce continuous text, it may become a new standard tool for aphasia patients in the future. Hopefully, we’ll see this technology being improved and utilized in the years to come.
