A tidy summary that differentiates simple “rule-based” chatbots, which match keywords in your text and respond, from serious artificial intelligence software, which analyzes your words in a more complex way to deduce what you really want — or better yet, what you really need.
10 Links to Learn About Chatbots in an Hour
Paul Boutin
393

Rule-based vs Serious Artificial Intelligence?

Hi Paul,

I believe that distinguishing rule-based engines from serious artificial intelligence systems is a wrong/misleading opposition, and I disagree too with the answers on the “tidy summary” link you mention.

So it depends on what you define with rule-based and serious artificial intelligence.

For me rule-based doesn’t means keywords matching (and so command-based bots), but it means an algorithm that process user utterance through a prededined set of rules, doing a intent guessing and meaning matching.

An example of a rule-based engine in this way is great opensource ChatScript, but following your statement:

serious artificial intelligence software, which analyzes your words in a more complex way to deduce what you really want — or better yet, what you really need.

ChatScript (that’s a famous rule-based chatbots engine), with his powerful natural language processing (NLP) pipeline, fit exactly the above mentioned description.

What I want to state here is that there is a too frequent but misleading antithesis between rule-based and mythical AI systems. In facts it appears to me that too often, people talk about AI, just referring to some sort of machine learning (for NLP) algorithms, but, in my opinion, in general, not all machine learning is artificial intelligence.

In facts, I admit:

I do not know what Artificial Intelligence means in 2017! And I suspect no one really knows…

With respect
giorgio