This cutting-edge technology, initially crafted to assist those with physical limitations, has blossomed into an unexpected marvel of versatility. Tailored for individuals grappling with conditions like Amyotrophic Lateral Sclerosis (ALS), "Silently Controlled Language Generation" transcends its initial function. Even individuals without physical impairments are uncovering the immense value of using biosignals for seamless communication, especially during dynamic activities like phone calls. "Silently Controlled Language Generation" user-friendly interface ensures that every interaction propels us toward a more inclusive future.
Now, let's dive into the project, a venture led by Unbabel. Unbabel is on a mission to create software and hardware that transforms communication for individuals navigating challenges in normal interaction. The crux of the innovation lies in blending artificial intelligence with biosignals. The role of IT is pivotal, contributing to the creation of a mobile application tailored for individuals struggling with conventional communication tools due to physical limitations.
Picture this: the software monitors incoming text messages, processes them, and identifies the context. Someone asks about scheduling a conversation. The model suggests personalized response options based on the user's personality and the received message. The user then effortlessly chooses a response aligned with their style. The application also offers vocalization of these options, allowing users to use facial muscle signals and Electromyography (EMG) to select or confirm their chosen response.
A standout feature of the project is the exploration of brain signals, particularly Electroencephalography (EEG). While EEG was the initial experimentation, the team encountered limitations such as signal noise and stability. Addressing concerns about user exposure associated with electrodes on the scalp, our approach steers clear of invasive alternatives like Elon Musk's NeuroLink, focusing on capturing muscle activation signals in a less intrusive manner.
The success story of the project is demonstrated in Unbabel's rapid implementation of the system, showcased in public spaces with an engaged audience. Witness Vasco (from Unbabel), a modern-day Houdini, as he seamlessly responds to multiple people without detection. This project doesn't just address specific physical limitations; it unfolds new avenues for nuanced and seamless communication, showcasing the potential of responsible AI to elevate human interaction to new heights.
This project is part of the Center for Responsible AI (NextGen) big project, which we will talk about in the next issue of our newsletter. Stay tuned!
Check the press report on this project: