Artificial emotional intelligence is almost here. Chat bots, as we’ve discussed, are displaying empathic skills that are getting their developers into legal hot water for encouraging, empathically, murder and suicide.
Now an AI voice is being introduced that not only speaks conversationally in multiple languages but that also listens carefully to you and mimics your personal style of speaking–the tone, pacing and emotional cues it hears.
Albert Mehrabian claimed in the 60s that effective communication is comprised of as little as 7% words and as much as 93% tone, speed and other vocal qualities along with body language and facial expressions (which can reveal emotional states). While his findings are often misstated, they are meant to serve as a guide when there is some ambiguity among the elements as to the message being conveyed. Research confirms that in an ambiguous situation, people rely more for meaning on those emotion-carrying cues like tone and facial expression than on whatever the words claim to express. It is those nonverbal cues that they trust.
Thus the introduction of Inworld’s new cloning voice–Realtime TTS-2–hits a further benchmark in making AI emotionally intelligent. The voice can hear your tone, pace and emotional cues and then mimic those elements in its responses to you and can do the same with others it communicates with. Not only is that emotionally perceptive and empathic, but that simple strategy of mimicking someone’s communication style, called linguistic mirroring, has been found to make the message more persuasive, building rapport, trust, and comfort.
For example, mimicking the opponent’s style in negotiation is a powerful tool that has been calculated to create a significant improvement in rapport with the other side. In one study of lawyers mirroring both writing styles and verbal styles of judges, legal teams who mirrored a judge’s preferred verbal style saw their chances of winning more than double.
The era of emotionally intelligent AI is coming and this voice cloning technology that is able to capture the emotional components of communication–tone, pace and other emotional cues–is likely to be more persuasive to you and others who are conversing with it. This is the beginning of AI-generated trust. A development to be watched.