Since 2022, when generative artificial intelligence took the world by storm, the technology has become the talk of the town. Not only does large language model (LLM)-based AI have exponential growth in development and application, but so have all other areas of the field. Additionally, artificial intelligence (AI) has moved from the digital to the actual world, improving people’s lives. It has affected every aspect of civilization, from early illness detection to language translation systems.
A similar problem for those with speech and hearing problems is being addressed by the Pune-based Indian startup Glovatrix. In the world, 1.5 billion individuals have some form of hearing loss, according to a World Health Organization (WHO) report. Although speech disability data is not kept by the organization, it is probably in the millions as well. These individuals frequently rely on written communication or sign language as their primary means of interaction. With the former, communication can be extremely slow; with the latter, the listener needs an interpreter available to convert sign language into spoken words.
This can frequently lead to problems finding employment, particularly in customer-facing or highly communicative positions. However, this is the role that businesses like Glovatrix play. The Indian business is developing the Fifth Sense, an AI-powered tool that can translate motions from sign language to voice virtually instantly and support independent, effective communication for anyone with linguistic difficulties.
The Indian AI startup Glovatrix
Co-founders Aishwarya Karnataki and Parikshit Sohoni established Glovatrix in 2021, and they each have a joint CEO position inside the business. In 2009, Karnataki encountered a youngster who had hearing loss. This encounter motivated her to study sign language and sowed the seed for her desire to help others who face a similar destiny. When she met the other co-founder, Sohoni, a data scientist with vast experience in predictive analysis who had gone through similar challenges in his own family, the idea blossomed into Fifth Sense. Sohoni was able to immediately connect with the cause.
Together with an engineer, the two have been operating out of a shared office space in Baner, Pune, forming a team of three. “Our vision is to facilitate seamless communication between people of all abilities and give every deaf and speech impaired person the ability to express and be heard in their natural language – sign language,” stated Karnataki in her explanation of her vision.
AI-powered gloves for speech-impaired people called Fifth Sense
The AI-powered gadget resembles a glove with a timepiece atop from a form-factor standpoint. According to Sohoni, the AI gloves are constructed of a lightweight fabric that is comfortable to wear for up to eight hours at a time. To operate a smartphone or perform other chores that call for a better grip, the user can stick their fingers out of the opening at the top of the gloves. The fabric has several sensors and a smartwatch built in to record any movements. It is possible to wash the cloth separately and it is detachable.
When it comes to the hardware, Sohoni clarified that the company sources the various parts it needs from other nations and then has them built independently according to its own internal design. The majority of wearable businesses in India have embraced this common format.
However, Glovatrix’s innovation has been found in its software. This system has two components that work together to provide smooth two-way communication. The AI-powered gadget itself is the first, and an accompanying app is the second. “We developed the AI architecture entirely in-house because there was no reference for us to look into,” Sohoni stated. It’s interesting to note that Glovatrix’s gesture-to-speech interface combines machine learning with various analysis techniques rather than generative AI.
The companion software records a gesture, turns it into audio, and plays it back to the user. When a person with hearing loss needs to listen to a speaker without using sign language, it also serves as a receiver. After the software has listened to sound, it translates it into text that the user can read. It’s interesting to note that the software can detect noises other than spoken words, like the doorbell ringing. Additionally, the device’s manufacturer states that it allows for smooth communication by operating practically instantly.
How Glovatrix is addressing the issues with connectivity and privacy
These kinds of smart devices frequently have privacy and connectivity issues. The majority of smart devices—especially those with AI—process and compute data on servers. This indicates that lag-free performance depends on a fast, functional internet connection. In a similar vein, in order for smart gadgets to work, a large amount of user data must be collected. In the event of a breach, maintaining this data on the server may also give rise to privacy problems.
Glovatrix has identified fixes for both issues. Any user-side audio that is gathered and converted to text is handled entirely on the device by the receiver portion of the app. Because it doesn’t require an active internet connection, this component is also lag-free. Regarding gesture-to-speech, Sohoni stated that in order to solve the connectivity issue, some key terms and individual alphabets will also be incorporated to the app itself. All of the remaining work will be done on the cloud, which necessitates a consistent internet connection, because AI models do demand powerful computational processing. Interestingly, the company’s cloud was developed locally and ought to aid in improving server-device connectivity for users in India.
It’s important to keep in mind that although the AI was trained in Indian sign language and can translate text between Hindi and Marathi, it can also produce audio in English and the majority of regional Indian languages using a text translation feature within the app.
Determining the product-market fit
As of right now, Fifth Sense is a prototype, but Sohoni disclosed that the company is about to start its first pilot test. The business is still optimistic that, in the upcoming six months, it will be able to determine the product-market fit.
Although the product is not yet ready for the market, Glovatrix has already observed the benefits of its gadget. It states that after utilizing Fifth Sense to converse with his interviewer, a person with speech impairment was able to secure employment.
And how much would the thing cost exactly? According to Sohoni, the goal of the AI gloves is to maintain a competitive pricing point to enable wide market adoption. He suggested it might cost as much as a mid-range smartphone, but he did not specify a price range. Glovatrix is also investigating a subscription-based revenue model as a means of lowering the price point and lessening the financial load on the final customer.
