Ask Siri whether you should believe in God and you’ll receive the reply, “I would ask that you address your spiritual questions to someone more qualified to comment. Ideally, a human.” Or, “It’s all a mystery to me.” Or the too-cute, “I believe in the separation of Silicon and spirit.”
Not surprising answers, given the nature of the question. Why should Apple put itself in the middle of a centuries-old discussion?
Apple’s Siri is a virtual intelligent assistant, but there is nothing intelligent about Siri or its peers, such as Microsoft’s Cortana, Google’s Assistant and Amazon’s Alexa. Virtual assistants are not conscious and do not “think” like humans. Their machine-learning software has enabled them to provide greater high-value content to humans. They are getting better at understanding the idiosyncrasies of people’s voices, accents and colloquial speech. A person can carry out a conversation with these devices giving the appearance of intelligence and consciousness.
There are large customized versions of intelligent assistants found in call centers that interact with customers unaware of the artificial voice on the other end. Another version involves virtual assistants that can take facts and write articles much like a reporter.
Darn near human
The reason these virtual assistants are getting better is because the content they collect from all these questions is used to improve the machine-learning software performance. The more content the virtual assistant acquires, the more complex answers it can handle.
Specifically, advances in machine learning and sensors have allowed for the development of “smart content.” Smart content is derived from virtual assistants, sensors collecting data from millions of conversations, thousands of internet sources, and sensors collecting data on physical surroundings. The machine-learning software determines the meaning of the data and creates the smart content verbalized by the virtual assistants.
The best example of smart content is robo-advisers, a combination of tremendous amounts of data and machine-learning software that use this data to make investment decisions for financial portfolios. Other examples can be found in our cars, as on-board software makes decisions thousands of times a second to ensure a smooth and safe ride. Autopilots on airplanes and automated farm machinery do the same thing.
Soon, smart content will be in an emergency room, making decisions on treatments faster than the ER staff.
Read Carone's entire commentary on the Chicago Tribune website.