Unveiled at CES 2024, the hyper-realistic avatar preserves the voice and persona of a 24-year-old lady with ALS to allow unprecedented communication and connection, complemented by a private, on-device AI from Lenovo to optimize textual content prediction and output
At CES 2024, international expertise chief Lenovo, assistive expertise non-profit the Scott-Morgan Basis, and AI-generated video innovator DeepBrain AI revealed a hyper-realistic AI avatar devoted to preserving the voice, persona, and bodily mannerisms of somebody with a degenerative illness. The groundbreaking avatar, created with DeepBrain AI expertise and spearheaded by Lenovo, opens new pathways for making use of generative AI to accessibility challenges and helps advance Lenovo’s imaginative and prescient of AI for all.
As a part of an ongoing collaboration with the Scott-Morgan Basis, Lenovo proposed creating an AI-powered avatar to leap previous conventional voice synthesis and remodel communication and connection for individuals with extreme disabilities. The Basis invited Erin Taylor, a 24-year-old lady just lately identified with amyotrophic lateral sclerosis (ALS), a neurodegenerative dysfunction that usually results in full-body paralysis, to take part and assist pilot a proof of idea for assistive communication.
Lenovo then approached DeepBrain AI, a pacesetter in next-gen avatars and a Lenovo AI Innovator companion, to make use of its generative AI expertise. Sponsored by Lenovo, the DeepBrain AI crew captured an in depth full-body video of Taylor in its California studio. The video turned the foundational knowledge for her new AI avatar, debuted at CES 2024, and providing 96 p.c true-to-life accuracy.
“This can be a improbable instance of the untapped potential of generative AI to empower people and construct a extra inclusive future, particularly with such visionary collaborators,” stated Scott Tease, vice chairman and common supervisor of Excessive-Efficiency Computing and AI at Lenovo. “Leveraging our expansive options portfolio and in-house experience, we love discovering these alternatives to reimagine what’s potential and assist
Lenovo’s additionally developed a prototype for a brand new predictive AI software to speed up textual content enter when a standard keyboard or voice-to-speech expertise will not be an choice—as with superior ALS. The non-public, on-device AI operates solely offline, leveraging a compressed massive language mannequin (LLM) to supply the velocity, reliability, and accuracy important when speaking well being info or different pressing wants. This follows Lenovo’s method to Hybrid AI, working throughout public, non-public, and private basis fashions that the corporate unveiled at its annual Tech World occasion in October 2023.
In contrast to frequent predictive textual content engines that present a single suggestion, this novel AI resolution supplies a number of choices after every character enter to higher help customers.
The Scott-Morgan Basis and Lenovo AI consultants are working to combine the predictive AI into the backend of their built-in communication platform showcased at CES. The platform combines the avatar, an unique round keyboard optimized for eye-gaze monitoring, and multimodal inputs to assist a variety of bodily capabilities.
“Our imaginative and prescient is constructed on a dedication to making sure a common proper to thrive by expertise,” stated Andrew Morgan, chief govt officer of the Scott-Morgan Basis. “AI is accelerating work we started years in the past, and our sensible collaborators are serving to us utterly change the accessibility panorama. Collectively, we’re lighting a beacon for what’s potential—not only for individuals with ALS, however for everybody.”
The DeepBrain AI avatar expertise was initially developed to energy generative AI people or assistants—such because the one showcased at a current Lenovo Components 1 occasion—or to create versatile “twins” of celebrities, sports activities figures, or different influencers. Right here, DeepBrain acknowledged an opportunity to deploy its proprietary expertise as assistive tech.
DeepBrain AI gave Taylor complementary entry to its AI Studios, empowering her to create an almost limitless stream of digital content material.
“We’re continually amazed at how our companions and shoppers consider new methods to make use of AI-generated video. Lenovo had the foresight to share their imaginative and prescient on how this expertise may very well be applied as a element of an assistive expertise platform,” stated Eric Jang, CEO of DeepBrain AI. “Future developments at DeepBrain will allow hyper-realistic avatars for everybody.”
After a short collection of escalating signs, Taylor was identified with ALS in the summertime of 2023. Her profession in botany, forestry, and associated work within the wild was abruptly paused as she and her mom confronted numerous questions concerning the development of the illness. The shock was particularly potent given the illness primarily impacts individuals over the age of 40.
“Feeling so devastated and misplaced, we have been grateful to the Scott-Morgan Basis for inviting us to assist construct one thing new and form the longer term,” stated Taylor, who lives outdoors Los Angeles along with her mom, Lily. “The avatar is unbelievable, however what issues most to me is that this expertise may assist change the lives of many extra individuals going through ALS or different illnesses—a method of displaying how vivid the longer term can nonetheless be.”
“I’ve little question this pioneering proof of idea will sooner or later assist form much more staggering improvements,” Morgan stated. “Our basis was constructed on the dream that expertise, and AI particularly, can break by extreme disabilities and guarantee a proper to create, join, and stay life joyfully. We’re making that occur.”
Study extra about Lenovo’s imaginative and prescient of AI for all, the Scott-Morgan Basis’s transformational work, and DeepBrain AI’s avatar expertise.
