Smart speaker

A smart speaker is a type of speaker with voice command-enabled services and an integrated virtual assistant to offer hands-free activation. Smart speakers can act as smart devices with wi-fi and bluetooth integration, allowing integration with the Internet-of-Things.

Smart speaker use has increased steadily, with an NPR and Edison Research study finding 24% of the U.S. population over 18 owns a smart speaker.

While smart speaker technology has evolved, consumer acceptance of the devices has risen, with research showing perceived ease of use, quality and enjoyment using the system strongly affects the acceptance of the speakers.

While a significant portion of the population own smart speakers, ownership varies between demographics. Pew Research found that adults younger than 50 are more likely to own smart speakers, and 34% of Americans earning over $75,000 USD per year own a device compared to only 15% for Americans earning less than $30,000 USD per year.

Voice recognition technology has existed before smart speakers and voice assistants. IBM released a tool called Shoebox at the 1962 Seattle World’s Fair, that was able to recognize 16 spoken words and digits and perform mathematical functions. Scientists at Carnegie Mellon University developed a technology called Harpy in the 1970’s that could recognize over 1,000 words for the U.S. Department of Defense and its Defense Advanced Research Projects Agency (DARPA).

Smart speakers use voice recognition and artificial intelligence software in order to understand, analyze and respond to user requests. Before a device can respond to a command, the information has to be cleaned and labeled for the algorithm to interpret. This takes enormous amounts of linguistic data, including differences in language, dialect and environment. After the voice data has been cleaned for use, the artificial intelligence must determine context of the question or command before determining an answer. Due to the large amount of smart speakers in the marketplace, there is a huge amount of data for the algorithm to learn from which has resulted in voice assistants’ performance to learn and perform better in a shorter time frame.

Smart speakers have raised privacy concerns by consumers wary of the continuously listening devices, and how their data will be used by the technology companies running the smart networks. While Amazon and other companies have stated the devices only listen to commands following “hot words”, independent studies have found 30-38% of audio recordings by Alexa Echo Dots were human conversations, showing the devices capture more audio than just commands following the hot words.

Wiretaps are another concern, with British security researchers showing the devices are capable of being infiltrated by government intelligent services.

Smart speakers come in all shapes, sizes and price ranges depending on the software included such as the voice assistant and control center capabilities, or the quality of hardware. While there is a limited number of voice assistants currently on the mainstream market, there is a much larger range of smart speaker manufacturers. While companies like Google, Amazon and Apple utilize proprietary voice assistants, other smart speaker manufacturers affiliate and integrate with existing voice assistants.

Voice assistants are digital software platforms capable of voice recognition, natural language process and speech, via a smart device in order to offer services such as scheduling appointments, sending messages, playing music, calling and more. Voice assistants sometimes act as a liaison or hub to allow multiple smart devices to connect, enabling a wide array of separate technology and devices to be controlled from a person’s voice.

Voice-operated automation and customer support is a rising technology, due in part to the increase in smart speaker ownership and consumer willingness to interact with the technology. Smart speaker and voice assistant-based customer service allows companies to know much more information about the person calling such as their purchase history, profile, activity and demographic. It also allows the company to quickly respond and get back to the customer in a timely, less intrusive manner. Companies continue to work at customer service AI, in order to improve NLP, ensure quality service to a growing number of consumers.