AI Voice Assistant
AI Voice Assistant with Open Source Technologies
Voice-activated products allow users to outsource simple interactions with their devices, thereby reducing screen time and enhancing the ability to multitask. Existing products connect to cloud services that are known to collect private data of users. These products do not work without an Internet connection. How can a privacy respecting prototype with offline capabilities for core functions look like?
/ CEO OPNTEC
A requirement is a voice assistant that can work vendor or platform independent and avoids a technology lock-in.
In order to provide a solution that is vendor independent our approach is to provide the solution based on open source. Secondly, in order to provide platform independent "skills", the mini-apps for the functionalities of the virtual assistants are designed to be transferable between different installations.
Virtual assistant solutions and technologies are incompatible. Current platforms prevent service providers to easily transfer solutions from one platform to another. How can a platform provide this option actively?
There is no standardized skill language available across different platforms. We are working on a skill language that is transferable and provider independent.
An issue preventing many users from using voice assistants is the question of user privacy. Users do not have control when data is sent to cloud providers and are not able to see what kind of data is sent and stored on platform providers machines. There is no option to validate if privacy choices and policies are adhered to. The user only has the choice to either accept transferring their data or not.
Users should have full control over their privacy settings. Firstly the device needs to have options to completely switch off voice activation. Secondly the platform needs to be open source in order to enable the verification of privacy settings and features.
A full-featured virtual assistant needs to be able to integrate IoT devices.
Machine-to-machine communication between smart sensors and mini-PCs such as the Raspberry PI are perfect examples of ‘The Internet of Things’. For example, appliances in the home can be monitored and controlled wirelessly by a homeowner wherever they are. How can this be done through voice commands? In a prototype scenario it we showcased the possibility of an "off-cloud" assistant that connects to the Shelly switch and is able to switch on/off appliances, light and smart meters.
A full-featured prototype with a rich API was developed as part of this project. The prototype can be installed on a Raspberry PI or on desktop systems using Linux. Future opportunities are open with further developing the user setup interface and adapting it to other platforms.
Virtual Assistants at OPNTEC
OPNTEC builds voice assistants based on open source conversational frameworks developed in cooperation with FOSSASIA. The motivation behind projects like SUSI.AI is to create an ecosystem made up of conversational web services, AI technologies, search engines, smart devices and open source communities. Our projects are based on the principle of privacy and collaboration. It provides user freedom and enables user to have the complete control over his/her own data. The goal is a voice assistant that is capable of chat and voice interaction and by using APIs to perform actions such as music playback, making to-do lists, setting alarms, streaming podcasts, playing audio books, and providing weather, traffic, and other real time information. Additional functionalities can be added as console services using an external API. It is able to answer questions and depending on the context will ask for additional information in order to perform the desired outcome.
How our services bring about success
Misako I. (UNESCO)
I discovered OPNTEC services through my research about open source solutions on the Internet. We were looking for a project partner in the area of Artificial Intelligence. We are working together with OPNTEC for many years now.