The rise of the virtual assistant
I’ve tried a few of the so called ‘virtual assistants’ over the years, crude applications that run in the cloud, on my smartphone or both and they never fail to disappoint. I’m afraid my expectations are high, set by the character Jane, created by Orson Scott Card in the Ender series of books, many years ago. Scott Card’s track record in predicting the near future is second to none in my mind (I’m biased as I loved his books) but Jane is not yet with us. Siri I’m afraid doesn’t come close.
In fact according to Ray Kurzweil we won’t see a human like virtual assistant, or perhaps more accurately virtual companion until around 2029, by which time I will be 65 and definitely in need of some assistance if the current state of my memory is anything to go by. I already desperately need real-time assistance in face recognition for example, I’m now at the point that the only names that come to mind instantly are the ones I use daily, people I meet once a month are an effort, and it gets worse from there. My ability to work with ideas and concepts is fine though, but names of anything are a challenge.
With the release of the movie ‘Her’ though the idea of the virtual companion is likely to get a huge amount of attention and hopefully this will inspire developments in the field, in the same way that I was inspired nearly 30 years ago by Jane.
There’s some serious issues though with the idea of a virtual assistant, and it’s access to information. Our current model of information access segregates information into at least 2 domain’s, work life and personal life. For me this segregation is completely artificial. My work life is seamlessly blended with by personal life in so many ways. I don’t work in a block of 8 hours a day, I don’t work for 5 days a week, my reading mixes work and personal life, my conversations do too. My calendar, contacts and personal knowledge archive are a seamless mix, my personal and professional social networks are not neatly separated. I use the same devices for work and play. I do promote certain assets that I create into controlled corporate repositories, but these are very much the tip of the knowledge iceberg. Stu digs into this topic in his post Is Enterprise Software Ready for the Personal Digital Assistant?
The other major issue is the privacy concerns associated with virtual assistants that will need some way to gather information about my physical world, which means they will need to see and hear as well as interact with the internet of things. The push back against Google Glass is worrying, but I see this as a temporary phenomenon, because I agree with Steve Mann, when he says Wearable Technology as a Human Right
So how will a virtual assistant work, if it only see’s half my life, and worse the missing bits are randomly distributed? well the short answer is that it won’t. The early experiments like Siri show the way, they work with my personal calendar, contacts, email etc. Assistants will be so compelling that this will create an unstoppable pressure to find ways to allow the virtual assistants access to company information, most likely by publishing the corporate information into the native applications on our smartphones, which of course are replicated in the public cloud. This will happen because the benefits offset the risks and because corporates will increasingly trust the defences built into our smartphones (eg proximity and biometrics) and the public cloud (eg dual factor authentication, encryption and clever intrusion and data loss prevention diagnostics).
I used to be inspired, now I’m starting to get excited, it won’t be long.