This article was published on Re-Work by Diane Bedet, Global Summit Creator on November 11, 2016.
With the increasingly rapid technological advancements in natural language processing (NLP) and deep learning, development of virtual assistants and chatbots has exploded this year, and new applications are being explored everyday.
One area of increasing interest is the improvement of healthcare and medicine through the use of virtual assistants, from diagnosing patients and providing support between appointments, to delivering medication alerts and educational materials.
Cathy Pearl is Director of User Experience at Sense.ly, where they’ve created a virtual assistant (VA) nurse to help patients with chronic health conditions manage their illness, bridging the gap between patients and clinicians. At the Virtual Assistant Summit in San Francisco, Cathy will share expertise on a panel to explore virtual assistants in healthcare, with a focus on how this could impact human labour and existing medical services. I spoke with her ahead of the summit to learn more.
What do you feel are the leading factors enabling recent advancements & uptake of virtual assistants?
Vast improvements in the speech recognition technology, and more acceptance that virtual assistants are mainstream, with things like Siri and Amazon Echo.
What do you feel is essential to future progress?
Making sure that in the excitement of bot platforms and virtual assistants, we do not forget about good design principles. In addition, we should focus on domain-expert virtual assistants, which are much easier to make successful.
What developments can we expect to see in virtual assistants in the next 5 years?
I think virtual assistants will begin to take over more and more low-level tasks. In addition, they’ll begin to learn more about an individual’s preferences and habits, so rather than ask for the same information over and over, like a form on a website, they’ll remember things about you, and leverage that (in hopefully a good way!)
What do you feel are the biggest challenges for adopting virtual assistants to everyday life & industry?
Making them easy to use, and useful! There are a huge number of bots/assistants coming out right now, but many of them are too difficult to use. One of the biggest problems out there is discoverability: how does your user know what your bot can and cannot do? Nothing is more frustrating than a prompt such as “How may I help you?” when really you can only say a few things the bot or assistant will understand. Integration with appropriate apps/back-end systems. If your virtual assistant wants to help with your health, or your finances, or your fitness–if it can’t talk to the other important systems in your life, it’s not much good.We also need more off-the-shelf analytics, to understand how bots and assistants are doing after being released to the wild.
Are we ready for emotional artificial intelligence?
Yes, but we must be cautious. We already have emotional AI–people already ascribe emotions to their coffee makers, their cars, anything. Technology is emerging to detect emotions, but it’s still in early stages. And worse than not detecting emotion is detecting the wrong one. In addition, even if you can detect someone’s emotion, what will you do with it?
Perhaps you notice someone is angry–saying “You sound angry” is not always a winning solution.
I believe people are ready for their virtual assistant to express certain emotions, when appropriate. For example, saying “I’m sorry to hear that” if the person says they have a bad day can help people feel better. It’s not that we’re trying to fool people–it’s that people are willing to engage in these familiar ways, even with a bot.