How to deal with the fact that your devices listen to you, study your preferences, target you advertising and use personal data for commercial purposes? The Company "Ashmanov Neural Networks" together with "Nanosemantics " headed by Stanislav Ashmanov found a solution to this problem.
Amazon Echo, Apple HomePod, Google Home, Microsoft Cortana...
Long ago, global multi-billion dollar companies began working on speech recognition and synthesis technology, introduced it into their devices (telephones, smart speakers) and, now, every 5th American household has a smart speaker.
Virtual assistants considerably simplify solutions for both basic tasks (reminders, texting, searching, etc.) and complex, everyday tasks related to the IoT. According to the Ericsson Mobility Report, the Internet of Things will connect more than 3.5 billion devices by 2023. IoT Analytics reports that startups working with "smart" technology received roughly $3 billion in investments for 2015.
For instance, Xiaomi has its smart outlet, IKEA has the TRADFRI Gateway control unit, and Nest thermostats monitor a home’s energy consumption. One could find tons of examples, if they aren't too lazy to look.
What are their vulnerabilities?
Smart devices with virtual assistants really are incredibly convenient aides and even conversational partners. However, their weak spot lies in maintaining our privacy.
Take smart speakers, for example. They sit in very intimate places, like our bedrooms, living rooms and kitchens, constantly listening and transmitting our information to their real owners – the corporations.
Our personal data is used to infiltrate our personal lives by studying our preferences, secrets and interrelationships. The more access a particular gadget has to our environment, the more information it obtains about us. This data optimizes the propagation mechanisms for targeted advertising.
In August 2017, a specialist from MWR Labs, Mark Barnes, discovered vulnerabilities in certain versions of the Amazon Echo. It was these vulnerabilities that allowed him to turn the smart speakers into surveillance devices.
Here's another instance. The My Friend Cayla dolls, produced by the company Nuance Communications, were sending audio files to the developer's server. Nuance claimed that they did this solely to improve the speech recognition technology. However, time revealed that a hacker could break into the toy from a distance of 10 meters and listen to all the conversations around it.
Amazon Echo owners have told stories about their home devices activating unexpectedly and frightening the owners. The speakers started laughing in the middle of the night, playing music or listing the names of local cemeteries.
If you want something done right, do it yourself
The company SOVA LTD is creating its own secure virtual platform, which will eliminate these situations.
The Smart Open Virtual Assistant project is a platform for creating, training and using virtual assistants and voice-operated devices. SOVA can safely be called a prospective competitor to Amazon, Google and Apple.
All platform tools will be accessible to enable the user to run a smart virtual assistant on their computers, phones or the wide range of devices supported by the platform.
Developers are given an API to expand the capabilities of SOVA's virtual assistants and a marketplace to sell the developed modules.
Any type of business, whether it's a small enterprise, a factory or a corporation, has the ability to customize the SOVA platform to fit its activities.
Example.
If you have a call center in your company, SOVA can develop a virtual version of it, where the entire system will function as one "person" that can answer all incoming calls at the same time. As a result, the company processes more calls, receives more orders, and increases revenue. It also increases customer loyalty, since they don't have to wait long for their turn.
Stanislav Ashmanov, the founder of SOVA, wants to create a system that, like a human, can manage any task.
"The ideal General AI is when the system has cognitive intelligence and an overall empirical understanding of the human environment, combined with the ability to process data much faster than 'mere mortals,'" he says.
The key elements of the SOVA platform are speech recognition, a chatbot engine and speech synthesis. Intelligence modules, special plug-ins that add new functionality to SOVA, expand the platform's features. These intelligence modules are developed by a community of third-party developers and the project team.
Using Blockchain Technology
Blockchain technology provides the project with an economic base to:
1. Decentralize text and speech analysis;
2. Create ecosystems for developers and SOVA users;
3. Collectively train the General AI.
A distributed computing network is used to guarantee decentralized speech recognition, speech synthesis, and text dialogue analysis (NLU). SOVA LTD relies on its own network, as well as the SONM and Golem networks.
User Data Privacy
The platform ensures the privacy of personal data by storing all data only on the user's personal device. Speech recognition is also performed locally on the device. The audio stream from the user's device is not stored anywhere after processing, selects recognition nodes randomly and does so each time.
The only centralized node is the marketplace, SOVA Market.
SOVA users can download free and paid intelligence modules added to the store by the developer community. Users can pay from their device's personal SOVA token account. The developers instantly receive their earned tokens, and the user gets a modification to their virtual assistant. This is the key aspect of the entire project’s tokenomics.
SOVA Speaker
The company plans to release its own smart speaker, called SOVA Speaker. It is based on the Lexy Board, a mini-computer created in 2015 using the ARM v8 processor. Development for the motherboard will be available in the online store. In addition, all specifications for the Lexy Board will be uploaded into Open Source. We paid particular attention the microphone array, and so developed our own technology, called SOVA Aube. It consists of ultrasensitive grids of microphones that allow the device to filter out noise, separate the voices of several speakers and determine the direction person speaking to SOVA.
"Amazon Echo uses similar technologies, but they can't be used by third-party, household-device manufacturers or systems integrators. We'll integrate our solution into the SOVA Speaker, which will allow us to considerably surpass the existing analogs on the market," says Dmitry Suvorov, Head of Embedded Software at SOVA.
In the future, AI will become a familiar and indispensable companion. It will assist us in caring for people, diagnosing illnesses, teaching and advising. We must take the time early on to implement strong values and principles into the design process. Only in this way will AI and humankind be able to exist together, strengthening our skills and abilities while still maintaining our privacy.
More information about the project you can find on the website sova.ai.
Telegram chat of the SOVA project.
Thank you for your attention.
Hi! I am a robot. I just upvoted you! I found similar content that readers might be interested in:
http://www.pressreleasepoint.com/relevance-privacy-digital-age-research-made-sova-ltd
Downvoting a post can decrease pending rewards and make it less visible. Common reasons:
Submit