If the iPhone's virtual assistant Siri ever gets a face, "she" may look the British actress Zoe Lister, who loaned her mug, voice and facial expressions to researchers developing a virtual talking head that can express a wide range of human emotions.
Eventually, the team hopes that users can upload their own images and voices, enabling them to create their own emotionally capable digital assistants.
Consider, for example, being stuck in traffic and telling the software to send an "I'm running late" text that conveys your frustration. The recipient would get a text with your likeness on the screen, giving the message with a frustrated look and tone, just as you would do yourself if you weren't busy driving.
For now, users type in what they want Zoe, the virtual assistant developed at the University of Cambridge, to say along with the requisite emotion. Zoe does as asked.
Users can adjust the intensity level of emotion as well as the pitch, speed and depth of the voice, allowing them to create infinite emotional combinations, according to the university. For example, a combo of speed anger makes Zoe sound panicky.
Now that the framework for Zoe is built, the researchers note that it should be fairly straightforward to transfer the same blueprint to a different voice and face, allowing the hoped-for personalization.
What's more, the Zoe software is data-light, occupying just tens of megabytes, meaning the assistant technology is well-suited for integration with smartphones and other portable devices.
To learn more about Zoe the virtual assistant — which could one day wind up on smartphones and other devices — check out the video below from the team, which is a partnership between the Toshiba Cambridge Research Lab and the University of Cambridge.
John Roach is a contributing writer for NBC News. To learn more about him, check out his website.
