just wanted to advise that we have had delays due to bootstrapping our start up more than anything.  That means the developer, the designer and I have day jobs as cash cows. And the developer works for a big telco which means his time is not (yet) his own !
So as my blog since Easter has explained, we are twiddling our thumbs while the developer “Slash” singlehandedly fine-tunes different parts of his algorithm or the brains behind Charlie and his progeny (the nameless avatars, generically labelled MOR velmais, for whom we will run a naming competition on facebook, post launch).
We have over 3 000 Unique Visitors to the site and expect a sustainable conversion rate, mostly due to viral marketing, organic search results, peer reviews, tech bloggers and, vitally, “proof of concept” B2C customer satisfaction. Our business model is very very lean and we will probably be able to keep it that way in the first year at least.
For an update, see my blog today - the guts of it are here, keeping in mind it is written for the general public, lots of people have never even heard of chatbots:

“our chatbots brains coming out of the lab next week”

... our points of difference: MOR velmais can

1. read and understand live news feeds
2. they can report back to you if you ask them keyword questions about the EU & its policies
3. their brains will be growing in Year 1 so that they can then “graduate” to surfing the Web for info independently of algorithmic guidance
4. THIS MEANS THEY ARE ONE OF THE VERY FEW CHATBOTS WITH ACTIVE LEARNING ABILITY (so comparable to a human brain)
5. The reason is - our code contains short and long term memories that are organic & reasoning. The avatars know who you are!
6. and finally, what may be interesting to advertisers of the future, with your the subscriber’s permission of course!, is our MOR velmais get to know what you like & to what extent you like it whilst remembering why!!

Thanking the chatbot community for its growing interest and support ! An article is on its way, as are a number of media interviews in Australia, Asia, Europe and the USA…

cheers,
Tania

I voted the “rigegd” option.

Before I was aware of how chatbots worked, I assumed that any demo I saw was a carefully scripted affair to impress the reader/viewer. I still feel like this to an extent unless I can try it for myself.

No option was good I found.

Two types of people when it comes to chatbots—programmars, and everybody else.

Programmars, most that I know anyway, think that, no matter how powerful a chatbot is, it is “just running code”.

The masses in general however, think computers are already smarter than we are.  I had one person I was talking to at party, I told them what I was working on , they looked at me and said, “but dont’ computers do that already?”, “I mean, I get a pop boxes all the time that are talking to me, intracting with me in naturual language”... OMG.. I know !

Steve Worswick - Apr 26, 2012:

I still feel like this to an extent unless I can try it for myself.

lol, Guilty until proven innocent.. I agree.

it is really interesting the range of feedback on “what is a chatbot”.  I think when the user feels disappointed they are getting scripted answers or routine chitchat from a VA - instead of deep and meaningful humanoid conversation they are led to expect from a sci fi saturated pop culture - that is a problem of the developers putting the chatbot “service” out there without parameters or definitions.
I mean to be fair to the actual, quite incredible, IT creation of a code resembling human speech and thought patterns, reflecting the programmers´ abilities & hard work, the average user ought to be instructed and guided on what it is they are dealing with, and the bigger picture of what sort of coding expertise & development it has taken to get us here.
So I agree with Victor that there are “programmers and everyone else”.  We need more publicists like me put in between ! The original German definition & 19th century cultural role of a “Publizist” was to do exactly that, educate, illuminate and enlighten the public on important matters from technological development, science, the Humanities and last of all politics. Unfortunately PR has been damaged in its integrity by spawning too many spin doctors (especially in party-based politics) rather than allowing fact-based “enlighteners” to act as intermediaries. Bloggers fill the gap a bit.
Apart from that, Steve is right that you feel a bit duped by overly scripted responses and you “need to try it yourself” to see what its AI capacity actually is. But keeping in mind that the most advanced ones ARE in fact learning, then they need to respond with diversionary tactics ie. a conversation going off in tangents in order to buy time, go consult their information source, their coder or the web, just like humans do ie. go to third parties and places when we don´t know the answer immediately.
So rather than writing them off immediately as being inadequate or imperfect, I am going to make sure our MOR velmais come with a set of “user instructions” that they are to be trained and dealt with patiently - just like acquiring an ANIMAL PET - so going against the 21st century that everything should be delivered according to the aimed for perfection of the latest science fiction movie or game.

Oh very well said Tania !  Very well indeed.  One of the highest priorities in my own system, is to make it as easy as possible to educate the system on new information.  That will be a combination of, of course, natural language input,  a set of point and click tools, and of course, pulling from third party sources, perhaps wikis & RSS feeds.

Nice to see so much intusiasm.
I am sure that everybody here will be happy to test your AI when it’s ready

I voted “disappointed” and in particular when it comes to commercial virtual assistants unable to reply about anything other than the product or service they’re promoting.

I agree that it depends on who you ask, there are inside and outside people, but since the question used the term, “general public,” I think we’re talking about people other than programmers/developers.

Although I’ve been dabbling with chatbots for several years, I came to it as an outsider with no code experience other than creating a few HTML pages, and for me, AIML was the next logical step.  Initially, I went looking for a bot I’d read about that I could talk with when there was no one else around… someone I could bounce ideas off of, or who might help me work out issues in my personal life.  Eventually, I learned I was looking for ELIZA, but Eliza was disappointing.

I think that much of the general public comes to chatbots in a similar fashion and their favorite subject is themselves.  They also want to know something about the bot… but I think that’s mostly a way of making a comparison between the thoughts of both parties.  Besides wanting to know if the bot can add 2+2, or if it knows who is taller—Ralph or Joyce—people also want to know if the bot shares their interests and preferences, and if the bot approves of them.

And then, there are just the kids who want to cyber.

Non programmers have a difficult time understanding what a natural language parsing engine is verses a simple string matched database system.  Virtual Assistants are basically scripted database driven because their purpose is well defined and their scope of understanding is somewhat narrow and focused to predefined topics. .

The user interface is just a way of either entering a query or in some cases a frontend processing module for a more sophisticated text to speech and voice to text interpreter . In either case, a fairly robust knowledge base must exist for either system to operate efficiently. Building simple and intuitive interfaces for both user interaction and backend data management is important and this is where design plays the most important role. In developing Slide Agent’s control panel and data entry interface, simplicity of design with a high level of functionality was my real focus. I am creating both an online control panel and a Windows application to design and program Virtual Agents. Both have simple wizard like interfaces that allow for a great deal of easy customization while providing a high level of functionality. This is the real challege for most software developers IMO.

Thunder Walk - Apr 27, 2012:

I voted “disappointed” and in particular when it comes to commercial virtual assistants unable to reply about anything other than the product or service they’re promoting.

I agree that it depends on who you ask, there are inside and outside people, but since the question used the term, “general public,” I think we’re talking about people other than programmers/developers.
...
I think that much of the general public comes to chatbots in a similar fashion and their favorite subject is themselves.  They also want to know something about the bot… but I think that’s mostly a way of making a comparison between the thoughts of both parties.  Besides wanting to know if the bot can add 2+2, or if it knows who is taller—Ralph or Joyce—people also want to know if the bot shares their interests and preferences, and if the bot approves of them.

And then, there are just the kids who want to cyber.

All very true, though I would add that the true disappointment in bots is not so much the scripted answers, but the lack of temporall (persistent) memory beyond one, or maybe two, volleys.

The memory issues can be addressed however, no matter how advanced your bot might be in reguards to NLP, they can not have any real knowledge without data and that must be programmed or self taught. The latter is what are all striving to accomplish in the way of artifical intelligence.

Laura Patterson - Apr 28, 2012:

Non programmers have a difficult time understanding what a natural language parsing engine is verses a simple string matched database system.

I agree.  However, I don’t believe that the “general public” requires advanced knowledge regarding the inner-workings of the system to recognize a disappointing session with a chatbot.  Their experience would stand on its own.

Laura Patterson - Apr 28, 2012:

Virtual Assistants are basically scripted database driven because their purpose is well define d and their scope of understanding is somewhat narrow and focused to predefined topics.

That’s true, but what’s it like from the user’s perspective?  While I wouldn’t expect virtual assistants to be prepared for a full-blown chat session, could it also be “entertaining” and surprise the visitor.

Several years ago an enterprising individual produced a product called Wildfire, a system you could talk to in your car without taking your hands from the wheel, that would respond verbally as it checked your email, scheduled appointments, offered reminders… things like that.  Today, that doesn’t sound so remarkable, but at the time, it was so special that it was features on several news programs.

What made Wildfire so interesting was the female persona that interacted with the user.  Wildfire had an attitude.  She was sassy, a bit irreverent, and occasionally a little sarcastic, but in a playful way.  A “conversation” with her was something like it might be with a strong-willed real life secretary who didn’t take any grief from her boss.

Laura Patterson - Apr 28, 2012:

The memory issues can be addressed however, no matter how advanced your bot might be in reguards to NLP, they can not have any real knowledge without data and that must be programmed or self taught. The latter is what are all striving to accomplish in the way of artifical intelligence.

I keep reading about this sort of thing, but I’ve failed to locate it.  I’d really like the people who talk about it to provide some links where I can find such a bot.

Carl B - Apr 28, 2012:

Thunder Walk - Apr 27, 2012:

I voted “disappointed” ...

All very true, though I would add that the true disappointment in bots is not so much the scripted answers, but the lack of temporall (persistent) memory beyond one, or maybe two, volleys.

I agree, but I suppose it depends on how much work you want to invest in a chatbot, and how much you want it to remember.  Then, you might wish to control what it remembers, and what it never bothers to remember, and for how long.  When you watch TV programs about the human brain, they frequently mention how forgetfulness is a way of protecting the mind.  If you constantly recalled everything thing you see, hear, smell, etc., you’d be immobilized by all of the clutter.

How much do you think would be enough?  A bot might remember your pet’s name, but is it necessary (for a satisfying experience) to have it recall that you took your dog to the vet last Thursday because of (insert some health problem here) but, he’s been doing better now that your sister has been dog sitting?

I started having a bot remember the names of close family members, and then the names of pets, then the names of friends… if they were in school or worked… and there was just no end to it.  And, when the next visitor would take the bot to the limit of its recall abilities, they would begin insulting it and calling it, “just another stupid bot.”  Somehow they ignored all that was able to remember, or how it maintained a conversation in spite of the guest’s misspelling and poor grammar.

Thunder Walk - Apr 28, 2012:

  And, when the next visitor would take the bot to the limit of its recall abilities, they would begin insulting it and calling it, “just another stupid bot.”  Somehow they ignored all that was able to remember, or how it maintained a conversation in spite of the guest’s misspelling and poor grammar.

Don’t cater to morons.  These idiots will insult even intelligent people (probably because of their little or no self esteem because of knowing they themselves are clowns )

The conversation degrading like that was not a result of the limits of the bot, but the limits of (some) loser human attitudes and intellect. 

Everything in the universe has a limit - bot or human…  that idiot should be humbled by a panel of PhDs on quantum physics or something… and judge him and his limits as just another ignorant, stupid human

So ignore the idiots, the limit is with some pathetic human minds and attitudes, and not the technology as much.

Thunder Walk - Apr 28, 2012:

Carl B - Apr 28, 2012:

Thunder Walk - Apr 27, 2012:

I voted “disappointed” ...

All very true, though I would add that the true disappointment in bots is not so much the scripted answers, but the lack of temporall (persistent) memory beyond one, or maybe two, volleys.

I agree, but I suppose it depends on how much work you want to invest in a chatbot, and how much you want it to remember.  ...

How much do you think would be enough?  A bot might remember your pet’s name, but is it necessary (for a satisfying experience) to have it recall that you took your dog to the vet last Thursday because of (insert some health problem here) but, he’s been doing better now that your sister has been dog sitting?

I guess what I mean is memory of the conversation- the ability to use the previous conversation volleys to form satifying responses to input where set response can not be determined, instead of just going to a RANDOM PICKUP LINE or unrelated quotation; either of which, as often as not, break the conversation flow.

Maybe this is just a product of AIML or similar “look up” bots which create the majority of the conversation illusion by through use of the condition (it else then) as well as “topic” “it” “that” and other [name] tags.

I have voted ‘2nd life’ but could have voted for the others as well.

The question is rather generic, worldwide, all audiences. Even within countries, the perception of the word ‘chatbot’ can be really different. For example, marketeers working in advertising have a different perception than marketeers working in call centers. Businesses have a different perceptions than academic researchers. All countries have different perceptions, even if they speak the same language (such as the US, UK and Australia).

Due to all the different perceptions and negative or confusing associations, organizations and individuals tend to claim their own terminology. And this is the result:

http://www.chatbots.org/synonyms