Siri and the dawn of the era of intelligent agents


I have recently done a number of interviews on the implications of Apple’s voice assistant Siri. To me, it’s looking very much like Apple has once again brought a technology to market precisely when it is sufficiently mature to impress. Voice control and ‘intelligent assistants’ are far from new, but haven’t been widely used to date simply because they haven’t been good enough.

The launch of Siri a year after the company was acquired by Apple has allowed them to develop what was already excellent technology to the point of being ready for the mass market. As with a number of other Apple releases over the last years, Siri’s launch is changing people’s perceptions of what technology can do, and opening many minds to new possibilities. We always knew we would eventually be able to tell our machines what we wanted them to do and have them respond. That era has begun. Though of course it still has a long way to go.

Today’s Sydney Morning Herald features is a nice article titled Siri: can you help make my company better?. It runs through some interesting insights on the state of predictive modelling, and closes with some quotes from me:

Siri could also be doing tasks based on voice command, says futurist Ross Dawson.

“We are heading towards more natural speech recognition and response. But the more useful thing is better human-machine interfaces. We’re still using a mouse to interact, exactly what we were doing decades ago.

“We’re on the verge of a significant transition to better human-machine interaction so (machines) understand what we mean. And voice is a big part of that. There are also gestures, eye gaze, facial expressions and even thoughts. But voice is one of the most important, as it’s generally the most efficient and is the least effort for communication.”

When people control devices with speech, the power lies in the system understanding the person, says Dawson, “what they mean or what they want by making statements”.

In future more business applications may be able to answer verbal questions such as “what are my stock levels today?”, or “tell me how sales are going across the globe” more readily than the hours it takes to navigate and analyse myriad reports.

The quotes used don’t express the underlying ideas as well as I’d like, but they allude to the primary point that machines will understand our intent far better than before. This will be both through using a proliferation of additional channels to communicate with them, over and above the keyboards and mice we use today, and also by being able to interpret what we really want and act on that dramatically better than ever before.

If we want something, we should just be able to say so. And machines will understand us and obey.