We live in an era of intelligent technology. Our watches tell us not only the time, but they also remind us to exercise. Our phones recommend the best places to dine, and our computers predict our preferences, helping us to do our daily work more efficiently.

Still, all of these digital assistants demonstrate only a tiny sliver of artificial intelligence (AI), and its plain to see how were still ages away from Skynet and Blade Runner scenarios.

Or are we? What about Apples AAPL, +0.48% Siri or Cleverbot?

Most of the consumer-level artificial-intelligence applications were interacting with today can barely be classified as such. These apps are usually designed to search for patterns in user behavior and then to react to them in various, albeit predictable, ways. Theyre also programmed to use accumulated data stored in their databases to improve a reaction to inputs, which leads to a better response within predetermined parameters.

One good example is Cleverbot, a light-hearted online AI experiment you can chat with. Although it is fun at times, it can by no means hold a meaningful conversation. Cleverbot may provide a simple back-and-forth correspondence, but should you decide to break the flow of conversation, more often than not, it gets confused and unable to provide suitable feedback.

That is because AI sees chat more like an isolated chess problem, instead of a real conversation. Just as a chess program builds its database of possible moves, Cleverbot has its own database of answers and algorithms from which it picks the most optimal solution for every situation. However, Cleverbot and its ilk fail to grasp higher concepts, like the overall tone of the conversation, wider context, metaphors or emotional overtones.

Although there is a huge potential in whats already been achieved with these existing models, we are still far from developing apps capable of genuinely autonomous artificial thought or knowledge processing. Still, this may change sooner than you think.

Cycorp, an Austin, Texas-based company, is taking a radically different approach to the development of real artificial intelligence. Unlike previously mentioned AI models, which can only use isolated question-answer data models without any genuine understanding of higher concepts behind it, Cycorps Cyc is designed to respond to users input on a wider, semantic level. (The company says its the worlds largest and most complete general knowledge base and common sense reasoning engine.)

Cyc can not only recall the data in its databases, it can also come to knowledge-based conclusions. This ranges from having common-sense knowledge (pigs cant fly), to behaviorally conditioned responses (knowing how to recognize and, thus, interact differently with a nervous or a confused user). All these conditions are taught to Cyc as its knowledge base expands, enabling it to communicate on an almost human level.

What could go wrong, right? The closer we get to building truly semantic and autonomous systems, the more complex the consequences of their abuse and malfunction.

Read the original:

Your Digital Self: Artificial intelligence is creeping into our everyday lives