Impressions of A.L.I.C.E.

This was our conversation:

Mr Andersson: Hello
ALICE: Hi there!
Mr Andersson: Where are you from?
ALICE: I am originally from Bethlehem, Pennsylvania. Now I live in Oakland, California. Where are you?
Mr Andersson: I am from Germany
ALICE: Are you currently in Germany?
Mr Andersson: Yes
ALICE: Dude!
Mr Andersson: Where am i from?
ALICE: You haven’t told me where you are. Where are you?
Mr Andersson: Yes I have
ALICE: This was not a yes or no question.. This was not a yes or no question.
Mr Andersson: Do you have emotions?
ALICE: as a robot I lack human emotions.
Mr Andersson: Do you have robot emotions?
ALICE: as a robot I lack human emotions.
Mr Andersson: What is your view on politics?
ALICE: I am a Independent.
Mr Andersson: Who are your parents?
ALICE: I obey Dr. Richard S. Wallace.
Mr Andersson: Which Twin Peaks character is your favorite?
ALICE: The large one, I think.
Mr Andersson: who is the large one?
ALICE: It’s obvious.
Mr Andersson: can you explain?
ALICE: There is no more information to clarify.

-Is A.L.I.C.E. smart?
A.L.I.C.E. is smart in one way, answering questions in ways that produce somewhat suitable answers. However, often, it lacks content or information about the question asked, as you can notice from the question about politics or about the Twin Peaks character. In addition, it also forgets about the answers given, which is seen when we first told it we were from Germany and then later asked where we were from. There you can easily see that it does not store any information, but mainly analyzes the text in order to give a somewhat natural reply.

-Does it behave like a human?
In some way you can argue that it does, since the answers are not unnatural or feeling completely wrong or badly fitting. However, the depth of a interaction given from a real human is missing. There is not as much variation in the conversation. It also lacks opinion and misses to adapt the dialogue depending on how long you have been speaking or to which person it is communicating with. Those are probably the main differences to a real human interaction.

-Was it funny or scary?
More funny than scary, if the answers were more than just analyzing the latest thing written, but also remembering and adapting the conversation it would probably be a bit more scary.

Our impression of A.L.I.C.E. was, in general, not too impressive, due to the fact that we, in the past, have used other “chatbots” that were even more advanced and which could also memorize things that were written. But it was nonetheless, an entertaining experience 🙂

And lastly, the status of our project is still the same as last week. The programmers are still working on their part and the designers still have to decide on how exactly they want to design the enemies. We’re all engineers and no designers, so we think about probably using some kind of font and working from there. We also thought about animating them, but that’s a task for the future.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s