Abstract:Currently, many smart speakers, even social robots, appear on the market to help people's lives become more convenient. Usually, people use smart speakers to check their daily schedule or control home appliances in their house. Many social robots also include smart speakers. They have the common property of being used in voice control machines. Regardless of where the smart speaker is installed and used, when people start a conversation with voice equipment, a security or privacy risk is exposed. Hence, we want to build a speech recognition (SR) that contains the privacy identification information (PII) system in this paper. We call this the SR-PII system. We used a Google Artificial-Intelligence-Yourself (AIY) Voice Kit released from Google to build a simple, smart dialog speaker and included our SR-PII system. In our experiments, we test SR accuracy and the reliability of privacy settings in three environments (quiet, noise, and playing music). We also examine the cloud response and speaker response times during our experiments. The results show that the speaker response is approximately 3.74 s in the cloud environment and approximately 9.04 s from the speaker. We also showed the response accuracy of the speaker, which successfully prevented personal information with the SR-PII system in three environments. The speaker has a response mean time of approximately 8.86 s with 93% mean accuracy in a quiet room, approximately 9.18 s with 89% mean accuracy in a noisy environment, and approximately 9.62 s with 90% mean accuracy in an environment that plays music. We conclude that the SR-PII system can secure private information and that the most important factor affecting the response speed of the speaker is the network connection status. We hope that people can, through our experiments, have some guidelines in building social robots and installing the SR-PII system to protect users' personal identification information.

Creating and Implementing a Smart Speaker

A New Mmwave-Speech Multimodal Speech System for Voice User Interface

Smart speaker design and implementation with biometric authentication and advanced voice interaction capability

VIRTUAL ASSISTANT USING ARTIFICIAL INTELLIGENCE

Smart Assistants for Smart Living: Ideology and Mythology in AI-Powered Smart Speaker Advertising

The Role of Artificial Intelligence in Voice-Based Systems: A Comparative Analysis of SAARUSH

Artificial intelligence voice assistant and home automation

Voice Assistant Using Artificial Intelligence

Marketing via smart speakers: what should Alexa say?

Digital Ventriloquism: Giving Voice to Everyday Objects

Embedding Artificial Intelligence for Personal Voice Assistant Using NLP

NLP based AI Voice Assistant

Building a speech recognition system with privacy identification information based on Google Voice for social robots

Smart Speakers, the Next Frontier in Computational Health

A Framework for Smart Home System with Voice Control Using NLP Methods

Home Automation Using Smart AI Assistant

Virtual Bot Powered by Machine Learning and NLP Technologies: Emulating Human-Like Conversations through Speech-to-Text Conversions

Digital life assistant using automated speech recognition

VOICE BASED VIRTUAL ASSISTANT

A Literature Survey on SpeakSmart: AI – Enhanced language learning guide

Making AI 'Smart': Bridging AI and Cognitive Science