The field of Artificial Intelligence has passed through many phases and has suffered many triumphs and defeats. The “AI winter” refers to periods in the history of AI where there was almost no public interest in AI research and no funding either. This was inevitably caused to due to previous high expectations that fail to deliver.

To get a feel of what it’s like, imagine this. You get carried away by the claims of a certain speech recognition software and decide to try it out. You’re told that you can speak naturally and that you merely have to dictate into a microphone and the software will type out your sentence in words. You start to get visions of handless computing without needing to type. However, when you actually try it out you’re dismayed by the flaws in the technology. You can barely get a proper sentence out and your realize that the promises are false.

This experience will lead to you being skeptical of any such AI claims in the future. Either with regard to facial recognition, handwriting recognition, or translation into foreign languages. Such were the reasons for decreased public interest resulting in periods of “AI Winters” where some researchers were even afraid to include the words artificial intelligence in their research papers.

Over a period of time, this blog will attempt to take examples of previous approaches to certain AI problems and then compare them to what we have achieved so far. This will hopefully provide readers with a picture of far we have come…and how far we have to go.

If you’re interested in AI, you’ll be surprised how far they’ve come along till now. You can try them yourself with online chat bots. Don’t be too hard on them though – they’re good. But not that good. My favorite place to find online chat bots is http://testing.turinghub.com/. It features some of the best bots out there. I have a personal liking for Landru – the namesake from the “Star Trek” episode – Return of the Archons.

You’ll find that each chat bot has it’s own personality. For example, the one called Alice is well – bitchy almost! You’ll need to experiment with them to find out which one is your style. Just ensure you keep the conversation linear. If you talk about something and then come back to it later, the bots don’t seem to remember it and therefore reveal themselves as not human. But otherwise, they amaze you with their understanding of what you’re talking about.

In fact, in the aforementioned site, it asks if you think you were chatting with the computer or one of the site’s adminsitrators – something called the “Turing Test”. Turring suggested that the proof of real AI would be when a human was unable to identify whether he was talking to a robot or not. In recent test, chat bots were able to fool humans 30% of the time. That’s pretty good. Stay tuned. They’re getting better. And when they do, we’ll be right here to tell you abou tit.