Speech is at the root of most human interactions. Increasingly, speech is also the user interface of choice for interacting with technology and businesses. In fact, just about every corner of our daily lives is being impacted by the speech revolution. Take these statistics, for example:
- 33% of people plan to invest in a smart home device in the next three years, according to a Clutch survey
- According to Persistence Market Research, the global voice-directed warehousing solutions market is expected to gain over $1 billion between 2018 and 2026
- Voice shopping will reach $40 billion by 2022, according to research from OC&C Strategy Consultants
While this research may not seem related, the fact is, these data points make it clear that the way we live, work, and shop are changing. But despite what seems like a society ready to embrace voice, there are still many skeptics.
According to a new study, “The State of Conversational AI and Consumer Trust,” 73% of respondents say they are unlikely to trust an AI-powered voice assistant to make simple calls for them correctly. That same study found that 81% believe that AI-powered voice assistants need to declare they are robots before proceeding with a call. In other words, formal research is confirming what we all discovered anecdotally when the introduction of Google Duplex spurred a tweet-storm and countless think pieces about the ethics of having a conversational assistant do our bidding without informing us that we’re speaking to a bot.
We hear, quite often, that chatbots and conversational AI need to be more human. What I think that really means is that we want it to respond to us—and perform for us—at the same level as a human assistant. But we still want to know it’s a machine.
Would Google’s Duplex have freaked out so many people if it sounded like a robot? If its voice was stilted and mechanical, but it still managed to make a restaurant reservation, would it have raised so many red flags? I don’t think so.
People might want the voice assistants they willingly engage with—like Alexa or Siri—to sound more human, but they aren’t so sure that’s the case when the machine is the one initiating the interaction. If we want AI-powered voice assistants to declare themselves as such, wouldn’t it be easier to just give them a voice that makes it clear from the outset that we’re not talking to a human? Ultimately, the result is the same.
Theresa Cramer is the editor of Speech Technology. She can be reached at email@example.com.