Now that 2017 has officially started, we are looking forward to exploring the latest technologies and how they will impact our lives. It's clear that AI will infuse everything and it's expected to impact all industries including: healthcare, energy, transportation, and retail! One area expected to grow exponentially are virtual assistants. With the already established existence of Siri, Cortana and Google Now, we will see conversational agents becoming more and more personalized. The market is expected to grow drastically over the next few years and we'll be showcasing advances in virtual assistants and artificial intelligence from the world's leading innovators at the Virtual Assistant Summit on 26-27 January in San Francisco.
We asked some of the speakers presenting at the Summit for their predictions for virtual assistants in 2017. Here are their forecasts:
The term "virtual assistant" will be split into "virtual personal assistant" to describe general-purpose assistants like Siri or Cortana, and "chatbot" to describe specific enterprise-wide, multi-channel conversational agents. Although AI technologies like Deep Learning will continue to evolve and fulfill new goals, their success in Natural Language Processing will begin to stagnate and ultimately disappoint corporations and end users. On the other hand, applications that combine Machine Learning with complex lexicons will gain a competitive advantage.
By the end of 2017, we will see the appearance of the first "360-degrees chatbot" that can operate as online customer support, semantic website search, and e-commerce search. These chatbots will be the central focus of customer interactions and will reshape how websites and mobile apps are being designed.
In the afternoon of day 1, Jordi will be presenting on Building Chatbots with Personality to Connect with Customers.
We’ll see more examples of truly conversational virtual assistants, which means, they will have multiple ‘turns’ and remember what happened earlier in the conversation. They’ll remember more personal information, such as pre-existing health conditions, and store a history of your habits and preferences. They’ll learn when you like to be chatty, and when you don’t want to be bothered. People will also become more comfortable speaking to devices in their homes and cars.
We’ll see the ability to personalize based on different people within the same household, so different users of one device like an Amazon Echo or Google Home can have different responses. In addition, I predict a big surge in attempts to do emotion detection, and some spectacular failures in that arena. But it will lay the groundwork for a more empathetic and personable companion.
Cathy will be sharing her expertise during the 'Virtual Assistants in Healthcare: Will they Assist or Replace Humans?' panel session on day 2 of the summit.
Less hype, but also less stigma. People will begin to get a sense of the real world applications of AI. Because these are far more mundane than the AGI of science fiction, we’ll see fewer cartoonish accounts of robots taking over the world and eliminating humanity. We’ll also see the continued emergence of verticalized AI (highly specialized agents that do one job exceptionally well). There are maybe a dozen in market; I’d expect to see the beginnings of a real AI agent marketplace.
Dennis will be presenting on The Emergence of Bring Your Own Agent (BYOA), where Dennis explains how BYOA promises to provide a host of benefits for both employees and employers, and is set to change the nature of work.
Today's virtual assistants do not have real conversations with humans. Popular speech activated AIs like Apple Siri and Amazon Alexa merely answer questions or do things on our behalf. For example, we could say: "Hey Siri set my alarm for 6am," and Siri will do that immediately. Yet, neither Siri nor Alexa is able to handle the next thing we might say: "Thank you so much, but that's too early for me." This is because they are unable to navigate what transpires over a series of questions and requests.
One question gets one answer. One request gets one thing done. But if we try to discuss a topic over a number of exchanges, they will tell us that they don't understand what we have said. There's no context to help them form the setting of a dialogue. And without that, an actual conversation is not possible.
There is good news. With advances in artificial intelligence this year, we will begin to see bots that can carry on a conversation with us in natural language. This will set the stage to develop virtual assistants that can be truly useful.Jack will be presenting on the various conversational AI's datalog.ai have created, including the enablement of personalized telemedicine in neuroscience which is specifically directed at chronic sufferers whose cognitive capabilities are degrading over time.
I think we'll see a decent amount of the hype/novelty wear off around virtual assistants, which is a good thing. I am optimistic there will be a renewed focus on solving actual problems for customers rather than just building virtual assistants for the sake of technology. I think there will be new companies that emerge that are "bot first" as well as an influx of larger brands that begin to use virtual assistants as a more mainstream marketing channel. Both of these forces will help each other in getting consumers more comfortable and used to interacting with virtual assistants.Danny will be presenting in the morning of day 2, his topic will be on How I'm building an emotionally aware bot, a bot that can track mental health.
Other predictions for virtual assistants and artificial intelligence:
The Virtual Assistant Summit will also be running alongside the Deep Learning Summit, 26-27 January in San Francisco. Attendees can enjoy networking with additional AI experts as well as viewing sessions from this agenda as well.
Virtual Assistant will be explored at the following events: