AI think, therefore AI am
by Terry Walby, UK MD of IPsoft
Thursday, May 24, 2012
Last week saw the 2012 Loebner prize take place, the competition in which computers attempt to convince judges that they are human. A few years ago, the results of a competition to demonstrate a machine mimicking a human may have seemed inconsequential to our everyday lives, but today the story is becoming more and more relevant. Increasingly machines use chatbots to interact with users online, or push out discussions on twitter, and the introduction of Apple’s Siri to the consumer market last year made all of this technology much more visible.
For all the marketing fluff, a system that proves it can have a conversation like a human is not really the point. For example, what Siri attempts to do is not to interact, it is to respond and react to simple questions and commands, but as with the “voice recognition” which has been available in high end cars for several years, these are simple tasks which provide little real-world gains in efficiency. But if the principles of natural language interpretation and interactive dialogue could be combined with the ability to trigger the automated execution of tasks, then we would be closer to seeing tangible efficiencies through Artificial Intelligence (AI).
Efficiency is a word that all businesses are drawn to, and if businesses are to use technology to drive such efficiency, what better place to start than the IT department, where use of technology is second nature? If an expert system could be “taught” to act as a virtual service desk, surely that would release the potential of AI to add tangible business value? Imagine if you could deploy virtual service desk staff which are able to think, act and interact independently without human intervention. Contactable 24/7 by a phone, email, Instant Messenger or SMS, they would not only listen and understand what is required, but would have the ability to trigger automated actions, run tests and execute fixes. They would remember everything, creating a vast database of every action undertaken or issue ever resolved, which could be recalled instantly.
The systems behind this principle would not just be having a conversation, but be mimicking the behaviour of a team of service desk engineers. They would continue to learn from every problem they encountered, every request they received and every action they took, and share that knowledge between them. Also, an AI system can handle many tasks at once, can instantly correlate complex information, and can make decisions at lightning speed. But unlike the human engineers they mimic, our virtual service desk staff don’t go sick, don’t forget anything and never come in to work with a hangover. And they never ask for a pay rise!
If this all sounds too far removed from a piece of software pretending it’s a person, it’s not! The principle of a virtual service desk engineer may sound like science fiction, but in reality, the technology isn’t as far away as you might think. Innovative companies have been working on creating this for over a decade. Done right, technology such as this will shake the IT outsourcing industry to its core.