2.2 Utilizing speech recognition for automatic phone responses during COVID-19 
Similar to the example shown above (see section 2.1), babies start listening at an early age to their parents’ voices and recognize them. They can hear the difference from their voices and strangers’ voices, so they might find more comfort from their parents’ voices than a stranger’s voice. However, unlike humans, computers do not have ears that let them hear sounds, or brains to understand that the sounds are words or phrases. Therefore, they use sensors, like voice detectors (just like the human ears), to detect speech. Specifically, there are three main stages involved in speech recognition:
Preprocessing involves taking the speech sounds and turning it into something the computer can use. Specifically, when we speak, we create vibrations in the air. The computer then uses an analog-to-digital converter (ADC) to translate this analog wave into digital data that it can understand. To do this, it samples, or digitizes, the sound by taking precise measurements of the sound wave at frequent intervals (see Figure 10).
During the recognition stage, the computer must identify what has been said. In this context, it starts analyzing the data also known as waves (obtained from the first step) and comparing it with already stored data of other words to identify what the user said.
During the communication stage, the computer acts upon the translated input.
| Speech (/spitʃ/) Recognition (/ˌrɛkəgˈnɪʃən/) |
|---|
| Speech recognition develops methodologies and technologies that enable the recognition and translation of spoken language into text by computers. It allows, for instance, hands-free control of various devices and equipment and automatic voice translation. |
| ADC: Analog-to-Digital Converter (/kənˈvɜrtər/) |
|---|
| An electronic integrated circuit that translate analog signals into digital values for use in processing and control systems, for example, when using the voice recorder of our mobile phones, the ADC will translate our voices as analog signals to digital data that can be recognized, saved and processed by our mobile phones. |
Story 4: AI Voice Responders Help Call Centers During COVID-19 
An AI Voice Responder, as shown in Figure 11, provides a fast and cost-effective way for organizations to better manage various call center needs and maintain a high level of customer service during the pandemic. The AI-enabled solution works by quickly answering the customer's call without making the customer wait. It can also automatically classify calls and resolve common problems without any manual intervention, and classify calls based on subject and urgency, so companies can prioritize follow-up issues to solve more complex issues.
As the measures to deal with the new coronavirus pandemic have led to the temporary closure of some mission-critical call centers around the world, consumers are facing long waits as companies struggle to cope with the surge in call volume. This is where the artificial intelligence voice reactor comes to the rescue and diverts high call volumes and answers common questions, quickly solving common support problems without a live agent.
For example, the artificial intelligence outbound services of Baidu and iFlytek, as leading providers of artificial intelligence services, can benefit as call centers that are under tremendous pressure due to the novel coronavirus pandemic. At present, the company’s customer service dialogue AI platform can immediately solve problems over the phone, improve customer experiences and reduce costs.
Unlike previous generations of IVRs (Interactive Voice Responses), voice responders can understand natural language and enable callers to talk as if they were talking to a living agent. The IVR system accepts mixed voice phone input and key-key keyboard selection, and provides related responses in the form of voice, fax, call back, email and other contact methods. On the other hand, artificial intelligence voice responders are used to help users solve basic hardware problems of commercial products and replace the affected offshore call centers. In addition, voice AI solutions ask questions to identify problems, provide relevant solutions through text, and follow up if the problem is not resolved.
Story 5: Using psychological consultant robot during COVID-19
Two middle school students Thomas and Lee are talking the news released by Carnegie Mellon University (See Figure 12). "Did you hear that Carnegie Mellon University is working on COVID-19 detection in human voice?" "Ah! Isn't that possible?" "Don't put limits on the technology. Imagine that if it works, it's certainly better to reduce the risk of infection when people go out to get tested." “Yeah. Looking forward to it”. It determines the probability of a user being infected by analyzing his/her voice and comparing it with the voice characteristics of people who are infected with the COVID-19. Sure, until now this application is not a diagnostic system" and has not been approved by the FDA or CDC and therefore it should not be used as a substitute for medical examination by a health professional.
