[alicebot-aiethics] A.I. and anger

[Robby Garner]

Hi Andreas and everyone,

>
> I think that we should see these angery users as an opportunity
> to discuss some fundamental questions of A.I. and ALICE:

Yes, the predictions by people like Asimov and Turing, are now becoming
reality.

>
> 1) When there are problems in a conversation between
>     an A.I. and a human being - which one should be modified?

Only the AI can be modified.  The human being is a wildcard, the great
unknown factor, the unpredictable circumstance.

>
> 2) Could anger be the right reaction?
>      For example: If the level of frustration is too high?

It could be okay for the AI to become angry, but at this point the only harm
it can inflict is verbal abuse. So it depends on the botmaster's
sensibilities as to whether his bot should ever lose it's cool.

>
> 3) Are there any rules in conversation, which neither a human  nor an A.I.
>     should break, if she/he/it doesn´t want to cause frustration and
anger?

Be polite, and respect the other interlocuter. I must say however, that I
find it irritating when a bot is too polite. It comes off as being
condescending or something I guess.  I think bots with an attitude are okay,
and possibly more interesting to talk to.

As a botmaster, one can't expect the human to follow any rules. That is the
grreat challenge, and the thrill when you get it right and someone has a
good chat with your bot.

Cheers,

Robby.