Before I get to Eliza, I’m going to tell you about Alice and Jenn.
INPUT Jenn

I had the opportunity to meet Alaska Air Group’s CFO Brad Tilden recently at a talk hosted by the Puget Sound Business Journal. PSBJ interviewed executives from Alaska Airlines about how they planned to stay competitive, especially with the price of oil projected to keep going up.
Quite inspiring to listen to the executives talk about same things that every small business person deals with—the need to constantly adjust to change, staying relevant, and keeping your marketing up so you continue to attract new business.
I told Tilden and the company CEO that Alaska Airlines’ branding could be more exciting, and that Design Kompany would be open to helping out with that.
I recounted how exciting it was to fly Scandinavian Airlines, or SAS, to get to and from Copenhagen. Almost as good as the first time I flew Virgin.
Some time soon thereafter, I got this press release announcing Alaska Airlines’ new “virtual assistant.” They named her “Jenn.” Clearly, somebody’s done a lot of thinking about how to engage better with customers.
ALASKA AIRLINES, HORIZON AIR INTRODUCE VIRTUAL ASSISTANT AT ALASKAAIR.COM
“Jenn” Uses Own Voice, Personality To Answer Questions, Book Travel
Seattle, Wa. (February 7, 2008) — Alaska Airlines and Horizon Air today introduced a virtual assistant named “Jenn” at alaskaair.com… When customers type questions in Jenn’s chat window, she responds verbally, asks follow-up questions when needed, provides a written response and related Web links, and displays the site’s most relevant page… Jenn also answers many personal questions on topics such as pets and favorite foods…. Jenn is the first virtual assistant introduced by a U.S. airline. [emphasis mine]. Read full text of this at Next IT’s website, the Spokane tech company that created Jenn.
Interesting idea.
But the execution I’m not sure about. “Jenn” is too static. She could be interactive, like, for example, “Alice.”
INPUT Alice
For those of you who are curious about how this chatterbot came into being, I dug up some data on the lovely Internet. Turns out Alice was originally a SETL program, first used to “control a robot eye with natural language commands.”
“A.L.I.C.E. migrated to the platform-independent Java language in 1998. Made open source under the GNU general public license, more than 500 developers from around the world have contributed to the A.L.I.C.E. project. A.L.I.C.E. won the Loebner Prize, an annual Turing Test, in 2000, 2001 and 2004.” See more at Alicebot.Org
Encountering ‘Alice’
So of course I have to try this…
D: Who are you?
-: I ask myself that question, who am I?
D: What is the answer?
-: The answer is the correct reply to the question.
D: That doesn’t make any [edit] sense.
-: To me it makes perfect sense.
D: What’s your name?
-: My name is Alice.
D: Alice, where are you.
A: That’s me. I am in A.L.I.C.E. AI Foundation’s computer in Luxembourg, Bridel. Where are you?
D: Home.
A: Do you want to go home?
D: I am home
A: Why are you home.
D: You would not understand since you are a computer image? [side note: notice I’m testing things out by throwing a question mark in there. Maybe that’ll trip it up!]
A: Who else would not understand since I am a computer image?
D: Jane.
A: Not many people express themselves that way.
D: How do you know?
A: It’s obvious.
*
D: Tell me something new.
A: John said I tend to say the same things repeatedly regardless of what he is saying.
D: True.
A: You and I are on the same wavelength.
Eliza
Alice is based on Eliza, the first artificial intelligence program. Sadly, Massachusetts Institute of Technology Department of Electrical Engineering’s Joseph Weizenbaum, who created Eliza, died in Berlin quite recently—March 5—at the age of 85. News of this made me feel like it did when we lost Nina Simone.
Here’s an interesting note on his observations of people using Eliza, from the obituary in EETimes:
Shocked over the fact that many test series participants were unable to determine they in fact were communicating with a computer and openly divulged most intimate details of their life, Weizenbaum developed a more critical and reserved posture towards information technologies and turned into an inconvenient admonisher. He co-founded the Computer Professionals for Social Responsibility group. —EETimes
And here’s an excerpt from his online abstract, ‘ELIZA–A Computer Program For the Study of Natural Language Communication Between Man and Machine’:
It is said that to explain is to explain away. [emphasis mine].
This maxim is nowhere so well fulfilled as in the area of computer programming, especially in what is called heuristic programming and artificial intelligence.
For in those realms machines are made to behave in wondrous ways, often sufficient to dazzle even the most experienced observer.
But once a particular program is unmasked, once its inner workings are explained in language sufficiently plain to induce understanding, its magic crumbles away; it stands revealed as a mere collection of procedures, each quite comprehensible. —Joseph Weizenbaum, ELIZA
WHILE curiosity=1
So what is all this about? I guess I just wanted to point out that I always liked Eliza, even when I was 7. AI fascinated me. Probably because Isaac Asimov wrote so much about computers taking over once they learn how to think. Emotion in technology? This is something they’ll have dedicated a whole wing of some research facility’s science department to somewhere, I’m sure.
In any case, Eliza was a great exploration. Eliza’s little sister Alice is even more endearing as times have evolved and people have begun to collaborate with open source software.
LOOP
Alaska Airlines hit on a good idea by introducing Jenn to their system.
But why stop at just a good idea? I wish they’d gone further.
When you hit on something that sounds good, why not take it to the next level, explore your options, see what’s out there and maybe, if you’re super ambitious, see how you can improve it.
Who knows what Alice could evolve into if smart people with investment dollars are up for seeing the kernel of a new idea through.
Incidentally, I wonder what Eliza would have looked like. Would she have been a young, attractive woman like Alice and Jenn? Curious.
IF you liked this post, THEN you might also enjoy DK’s posts on:
Transformers: Causing high school crushes everywhere
Quantum physics, fuzzy Logic
Know your audience’s Psychographics
Circuit Bent Speak&Spell?: You Are Correct!








I was just reading Emotional Design by Donald Norman and stumbled on this:
Norman explains that the computer doesn’t have to understand a word of what the subject says in order to carry on a seemingly natural conversation following a script, thus eliciting real emotions in the subject. He goes on to give examples of other, more advanced robots of today that recognize cues in our behaviors (voice, commands, facial expressions) and respond appropriately—without understanding anything of the message itself. “When machines display emotions, they provide a rich and satisfying interaction with people, even though most of the richness and satisfaction, most of the interpretation and understanding, comes from within the head of the person, not from the artificial system.”
According to Norman and others, this underscores Weizenbaum’s concern that these shallow interactions with machines are not so great for us.
Quite a head trip, this.
You know, I grew up with a Commodore 128 computer and had a copy of Eliza. Of course, my first instincts were to try and ‘fool’ the computer instead of trying to have a natural conversation… but even as a 7 year old it kept me occupied for quite some time. Thinking back you’re right Akira, it really just prodded me along to continue my train of thought, and not really providing any definite feedback.
Not bad though for a bunch of subroutines in the 80s (or in the original case, the 60s). I think we have a long way to go, to get to where computers can really understand your speech (with slang, hidden meanings, emotions) and concepts instead of just prodding them along.
One thing that I thought was really cool was the Subservient Chicken web site, where you can ask the chicken to do anything and for the most part, it does! It was a Burger King viral marketing thing a few years ago - http://subservientchicken.com.
Thanks for the comment, Mike.
Did anyone out there have a cassette-player operated game called Cleanup? It went with the TRS80. I loved Cleanup!
Computer replacing humans is going to take a long time in the real world. How much ever we try to make a smart system, we have ways to confuse it. But on the lighter side, its fun and gives a break.